WarmTransfer
sbc trunking · published

SIP 503 and trunk out-of-service failures

Verified 2026-09-25 · 54 sources · tier 1–2

In SIP, a 503 Service Unavailable response indicates that a server is temporarily unable to process a request due to temporary overloading or maintenance 3. A SIP client receiving a 503 response should attempt to forward the request to an alternate server 8.

SIP 503 semantics and overload handling

When a client receives a 503 response containing a Retry-After header field, it should not forward any other requests to that server for the specified duration 47. At a SHOULD NOT strength, RFC 3261 directs proxies not to forward 503 responses upstream 46. Furthermore, RFC 3261 permits proxies to refuse the connection or drop the request outright instead of responding with 503 45. RFC 5390 observes that the conditions under which a server should issue a 503 are ambiguous, with implementations generating 503 responses for numerous reasons beyond actual overload 1.

The combination of 503 and Retry-After provides limited control and cannot prevent server overload or congestion collapse 2. Using 503 for overload can make traffic oscillate and shift between SIP servers, worsening the overload 4. To handle overload directly, RFC 7339 (Standards Track, September 2014) carries overload feedback hop by hop in Via header parameters oc, oc-algo, oc-validity and oc-seq; a client adds oc to the topmost Via it inserts and the server returns a value in the response's topmost Via 52. Under the default loss algorithm of RFC 7339, the oc value is a percentage from 0 to 100 by which the client must cut the requests it forwards to the overloaded server, and oc-validity is in milliseconds 51.

For server location mechanisms under RFC 3263, a failure is defined as receiving a 503 response, experiencing a transport failure (such as TCP connection failures or fatal UDP ICMP errors), or encountering a transaction timeout with no response 48. When such a failure occurs, the client should generate a new transaction with a fresh Via branch and send an otherwise identical request to the next entry in the DNS SRV-ordered list 49.

Diagnosing 503 responses

WarmTransfer's reading of the sources is that because proxies are told not to relay 503 upstream and Teams reports whether a 503 came from Microsoft (540998) or the SBC (560503), the first troubleshooting step for a 503 is to find which hop generated it before blaming the far-end carrier 43.

On Twilio Elastic SIP Trunking, termination capacity defaults to 1 call per second (CPS) per trunk per region, with self-service adjustments covering 1 to 5 CPS and higher allocations requiring Twilio sales 53. If the termination CPS threshold is breached, calls are rejected immediately rather than queued, generating debugger error 32001 for individual trunk limits or 32012 for pooled parent-account limits 54.

Platform behaviors and trunk out-of-service states

WarmTransfer's reading of the sources is that platform handling of 503 keepalive replies varies: Cisco Unified Communications Manager (Unified CM) flags the trunk as out of service, AudioCodes Mediant session border controllers (SBCs) allow failure codes to be configured, and Teams Direct Routing relies primarily on whether the SBC transmits OPTIONS rather than the specific response code 44.

Microsoft Teams Direct Routing

In Microsoft Teams Direct Routing, an SBC receiving a 503 with a Retry-After header in response to an INVITE must terminate the connection and attempt the next available Microsoft datacenter 24. A busy Direct Routing datacenter can return a Retry-After interval of 1 second 40. For outbound routing from Teams, Direct Routing defaults to failing over to another SBC in the voice route when an SBC returns final response codes 408, 503, or 504 33. Failover behavior can be adjusted for any non-2xx final code per gateway using the PowerShell command Set-CsOnlinePSTNGateway -FailoverResponseCodes 32. Direct Routing fails an outbound call over to another SBC only if the first SBC has sent no non-100 provisional response, to avoid double ringing 34. WarmTransfer's reading of the sources is that if an SBC issues a 180 or 183 response followed by a 503, Teams will not fail over to an alternate SBC despite 503 being a configured failover code 42.

For network-level errors—such as connection refusals or TLS timeouts—Direct Routing retries the same trunk from a separate Microsoft datacenter, which may reside in another geographic region 37. If declining a call produces multiple missed-call notifications, retry logic on the SBC or PSTN provider is misconfigured, and the SBC must be set to cease retrying when receiving a 603 response 28.

Direct Routing requires a successful OPTIONS exchange resulting in a 200 OK before call establishment is allowed 38. If an INVITE or OPTIONS sent to the Direct Routing SIP proxy contains an IP address in the Contact header instead of the SBC FQDN, the proxy rejects the message with 403 Forbidden 29. Teams considers an SBC healthy if OPTIONS messages have been received within the past 3 minutes, against a standard 1-minute interval 39. An SBC that misses this window is demoted rather than dropped: it will not receive primary routing priority, but it will still be attempted before call termination if healthy peers in the route fail 30. When multiple equal-priority SBCs are healthy, Direct Routing distributes outbound calls across them using a Fisher-Yates shuffle 35. SBC configurations for Microsoft 365, Office 365, and GCC environments should be configured with sip.pstnhub.microsoft.com, sip2.pstnhub.microsoft.com, and sip3.pstnhub.microsoft.com in that priority order, whereas GCC High and DoD rely on a single FQDN 36.

Direct Routing response code guidance applies specifically to Direct Routing PSTN traffic and excludes Calling Plan or Operator Connect architectures 25:

  • Direct Routing Microsoft response code 540998 with SIP 503 (Service state Inactive/DrainingTransactions) is expected while Microsoft SIP endpoints are down for maintenance and normally needs no action 26. If encountered repeatedly, Microsoft advises verifying that the SBC re-resolves Direct Routing FQDNs at least every 15 minutes 31.
  • Direct Routing Microsoft response code 560503 with SIP 503 means the 503 came from the SBC (maintenance or temporary overload); Microsoft says to check SBC logs, confirm the SBC is licensed for the concurrent sessions, and see whether the failures are tied to one destination country, region or corridor 27.
  • Direct Routing Microsoft response code 10320 with SIP 503 means a bot (for example built-in call recording) was unreachable or did not answer before the timeout 23.

Because Teams Call Analytics tracks calls only after they enter internal Microsoft infrastructure, troubleshooting SBC pairing faults and rejected INVITEs requires checking SBC logs directly, where Direct Routing delivers diagnostic text 41.

Cisco Unified Communications Manager and CUBE

In Unified CM, SIP OPTIONS Ping is enabled on the SIP Profile assigned to the trunk, and the matching Cisco Unified Border Element (CUBE) dial peer uses voice-class sip options-keepalive 15. Each Unified CM node running a SIP daemon pings each configured destination independently, forwarding active calls only to destinations confirmed reachable 14. A destination is marked out of service if it responds with 503 or 408, fails to respond to OPTIONS, or fails TCP connection establishment 20. The overall trunk is marked in service as long as at least one node receives a response other than 408 or 503 from at least one destination address 16.

When all remote peers fail, Unified CM marks the trunk status as "No service" and generates the RTMT alarm UC_CALLMANAGER-3-SIPTrunkOOS 22. Status Reason codes provide diagnostic detail:

  • local=1: No reply was received to OPTIONS pings after UDP retries 17.
  • local=2: Unified CM was unable to establish a socket connection to the destination 18.
  • local=3: DNS resolution failed due to an unreachable or misconfigured DNS server 19.

On CUBE running IOS XE 17.6 onwards, keepalive profiles default to an up-interval of 60 seconds, a down-interval of 30 seconds, a retry count of 5, and transport set to system 11. CUBE places a dial peer into a busyout state only when every destination inside its server group or DNS SRV profile experiences keepalive failures 9. However, if multiple dial peers share an identical keepalive profile and that ping fails, every dial peer referencing the profile is busied out 12.

AudioCodes Mediant SBC

On AudioCodes Mediant SBC, Proxy Keep-Alive functions via OPTIONS or REGISTER requests 7. The Proxy Keep-Alive Time parameter defaults to 60 seconds, while recovery is controlled by Success Detection Retries (default 1) and Success Detection Interval (default 10 seconds), which dictate the consecutive successful replies required to restore an offline proxy to service 7. The Keep-Alive Failure Responses setting permits administrators to specify up to 3 explicit SIP response codes that cause the SBC to classify the proxy as offline when returned in response to keepalives 5.

See also

Applicability

Cisco CUBE keepalive defaults and busyout rules apply to IOS XE 17.6 onwards 11​9. Microsoft Teams Direct Routing 503 response-code guidance applies to Direct Routing PSTN calls and not to Calling Plan or Operator Connect deployments 25.

What remains uncertain

Trunk out-of-service behavior across unstated vendor releases is not covered by the sources below.

See also

Related to

Sources

  1. 1
    RFC 5390 notes that the cases in which a server should send 503 are ambiguous and that implementations send 503 for many reasons, only some of which are actual overload.
    RFC 5390: Requirements for Management of Overload in the Session Initiation Protocol · Section 5.4 Ambiguous Usages · Checked 2026-09-25
  2. 2
    The 503 plus Retry-After mechanism is only a limited overload control: it cannot prevent overload of a SIP server or congestion collapse.
    RFC 6357: Design Considerations for Session Initiation Protocol (SIP) Overload Control · Section 1 Introduction · Checked 2026-09-25
  3. 3
    In SIP, 503 Service Unavailable means the server is temporarily unable to process the request because of temporary overloading or maintenance of the server.
    RFC 5390: Requirements for Management of Overload in the Session Initiation Protocol · Section 4 (quotation of RFC 3261 section 21.5.4) · Checked 2026-09-25
  4. 4
    Using 503 for overload can make traffic oscillate and shift between SIP servers, worsening the overload.
    RFC 7339: Session Initiation Protocol (SIP) Overload Control · Section 1 Introduction · Checked 2026-09-25
  5. 5
    On AudioCodes Mediant SBC, the Proxy Set parameter Keep-Alive Failure Responses lists up to three SIP response codes that, if received in reply to keep-alive messages, make the device treat the proxy as offline.
    Configuring Proxy Sets - Mediant Software SBC User's Manual Version 7.4 · Proxy Sets table, parameter 'Keep-Alive Failure Responses' · Checked 2026-09-25
  6. 6
    AudioCodes Proxy Hot Swap is disabled by default; when enabled, a response listed in the SBC Alternative Routing Reasons Set moves the request to the next proxy, and if no proxy is available the SIP message fails.
    Configuring Proxy Sets - Mediant Software SBC User's Manual Version 7.4 · Proxy Sets table, parameters 'Proxy Hot Swap' and 'Redundancy Mode' · Checked 2026-09-25
  7. 7
    AudioCodes Proxy Keep-Alive can use OPTIONS or REGISTER; Proxy Keep-Alive Time defaults to 60 seconds, and Success Detection Retries (default 1) with Success Detection Interval (default 10 seconds) set how many consecutive successes bring an offline proxy back online.
    Configuring Proxy Sets - Mediant Software SBC User's Manual Version 7.4 · Proxy Sets table, parameters 'Proxy Keep-Alive', 'Proxy Keep-Alive Time', 'Success Detection Retries', 'Success Detection Interval' · Checked 2026-09-25
  8. 8
    A SIP client (proxy or UAC) that receives a 503 SHOULD attempt to forward the request to an alternate server.
    RFC 5390: Requirements for Management of Overload in the Session Initiation Protocol · Section 4 (quotation of RFC 3261) · Checked 2026-09-25
  9. 9
    CUBE marks a dial peer busyout only when every destination in the dial peer (for example every target in its server group) has a heartbeat failure.
    Cisco Unified Border Element Configuration Guide - Cisco IOS XE 17.6 Onwards - SIP Trunk Monitoring · SIP Trunk Monitoring chapter, server group and DNS SRV subsections · Checked 2026-09-25
  10. 10
    For DNS SRV destinations, if CUBE gets a 503 or no response to an INVITE it marks that destination busyout and tries the next destination; the call is rejected if all destinations are busied out.
    Cisco Unified Border Element Configuration Guide - Cisco IOS XE 17.6 Onwards - SIP Trunk Monitoring · SIP Trunk Monitoring chapter, subsection 'OPTIONS Ping for DNS SRV Hosts' · Checked 2026-09-25
  11. 11
    On CUBE (IOS XE 17.6 onwards), a voice class sip-options-keepalive profile defaults to up-interval 60 seconds, down-interval 30 seconds, retry 5 and transport system.
    Cisco Unified Border Element Configuration Guide - Cisco IOS XE 17.6 Onwards - SIP Trunk Monitoring · SIP Trunk Monitoring chapter, keepalive profile defaults · Checked 2026-09-25
  12. 12
    When several CUBE dial peers share one OPTIONS keepalive profile and that ping fails, all of the associated dial peers are busied out.
    Cisco Unified Border Element Configuration Guide - Cisco IOS XE 17.6 Onwards - SIP Trunk Monitoring · SIP Trunk Monitoring chapter, SIP OOD OPTIONS ping group · Checked 2026-09-25
  13. 13
    In the DNS SRV case, Cisco's CUBE guide says CUBE monitors all destinations irrespective of the response it receives, naming 503 and 200 OK as examples.
    Cisco Unified Border Element Configuration Guide - Cisco IOS XE 17.6 Onwards - SIP Trunk Monitoring · SIP Trunk Monitoring chapter, subsection 'OPTIONS Ping for DNS SRV Hosts' · Checked 2026-09-25
  14. 14
    With OPTIONS Ping enabled, each Unified CM node running the trunk's SIP daemon periodically pings each destination and sends calls only to reachable destinations, instead of discovering failures per call through timeouts.
    Cisco Unified Communications Manager Trunks (Session Management Edition deployment guide 8.6(1)) · Subsection 'SIP OPTIONS Ping' · Checked 2026-09-25
  15. 15
    In Unified CM, OPTIONS Ping is enabled on the SIP Profile (Device > Device Settings > SIP Profile) assigned to the SIP trunk, and the CUBE side uses voice-class sip options-keepalive on the dial peer.
    Configure Options Ping Between CUCM and CUBE · Configure section (CUCM SIP Profile and CUBE dial-peer steps) · Checked 2026-09-25
  16. 16
    Unified CM considers the trunk in service when at least one node gets a response other than 408 or 503 from at least one destination address.
    Cisco Unified Communications Manager Trunks (Session Management Edition deployment guide 8.6(1)) · Subsection 'SIP OPTIONS Ping' · Checked 2026-09-25
  17. 17
    A Unified CM trunk Status Reason of local=1 means no response was received to OPTIONS after retries over UDP.
    Calls through Session Initiation Protocol (SIP) Trunk Failure · Scenario local=1 · Checked 2026-09-25
  18. 18
    A Unified CM trunk Status Reason of local=2 means Unified CM could not create a socket connection to the remote peer.
    Calls through Session Initiation Protocol (SIP) Trunk Failure · Scenario local=2 · Checked 2026-09-25
  19. 19
    A Unified CM trunk Status Reason of local=3 means DNS resolution failed (DNS server unreachable or DNS misconfigured).
    Calls through Session Initiation Protocol (SIP) Trunk Failure · Scenario local=3 · Checked 2026-09-25
  20. 20
    With SIP OPTIONS Ping enabled, Unified CM considers a trunk destination out of service if it does not respond to OPTIONS, responds 503 or 408, or a TCP connection cannot be established.
    Cisco Unified Communications Manager Trunks (Session Management Edition deployment guide 8.6(1)) · Subsection 'SIP OPTIONS Ping' · Checked 2026-09-25
  21. 21
    Cisco's CUCM-to-CUBE OPTIONS ping tech note says the OPTIONS messages are exchanged every 60 seconds by default.
    Configure Options Ping Between CUCM and CUBE · Verify / explanatory text following configuration · Checked 2026-09-25
  22. 22
    When all remote peers of a Unified CM SIP trunk are out of service, RTMT raises UC_CALLMANAGER-3-SIPTrunkOOS and the trunk page shows the status as No service.
    Calls through Session Initiation Protocol (SIP) Trunk Failure · Problem section · Checked 2026-09-25
  23. 23
    Direct Routing Microsoft response code 10320 with SIP 503 means a bot (for example built-in call recording) was unreachable or did not answer before the timeout.
    SIP 503 and Microsoft response codes - Microsoft Teams · Section '10320 503 Bot is unreachable or unable to answer before timeout' · Checked 2026-09-25
  24. 24
    In Teams Direct Routing, when the SBC gets a 503 with a Retry-After header in response to an INVITE, it must terminate that connection and try the next available Microsoft datacenter.
    Teams Phone System Direct Routing: SIP protocol · Section 'Retry-After' · Checked 2026-09-25
  25. 25
    Microsoft's SIP 503 response-code guidance applies to Direct Routing PSTN calls and not to Calling Plan or Operator Connect deployments.
    SIP 503 and Microsoft response codes - Microsoft Teams · Note at top of article · Checked 2026-09-25
  26. 26
    Direct Routing Microsoft response code 540998 with SIP 503 (Service state Inactive/DrainingTransactions) is expected while Microsoft SIP endpoints are down for maintenance and normally needs no action.
    SIP 503 and Microsoft response codes - Microsoft Teams · Section '540998 503 Service Unavailable - Service state: Inactive/DrainingTransactions' · Checked 2026-09-25
  27. 27
    Direct Routing Microsoft response code 560503 with SIP 503 means the 503 came from the SBC (maintenance or temporary overload); Microsoft says to check SBC logs, confirm the SBC is licensed for the concurrent sessions, and see whether the failures are tied to one destination country, region or corridor.
    SIP 503 and Microsoft response codes - Microsoft Teams · Section '560503 503 Service unavailable. SBC undergoing maintenance or temporarily overloaded' · Checked 2026-09-25
  28. 28
    If several missed calls appear for one declined call, the SBC or PSTN provider retry logic is misconfigured, and the SBC must be set to stop retrying on a 603 response.
    Teams Phone System Direct Routing: SIP protocol · Section 'Handling retries (603 response)' · Checked 2026-09-25
  29. 29
    An INVITE or OPTIONS to the Direct Routing SIP proxy whose Contact header uses an IP address instead of the SBC FQDN is refused with 403 Forbidden.
    Teams Phone System Direct Routing: SIP protocol · Section 'Detailed requirements for Contact header and Request-URI' > Contact header · Checked 2026-09-25
  30. 30
    An SBC that has not sent OPTIONS on schedule is demoted rather than removed: it is not tried first, but it is tried again before the call fails if the other SBCs in the route cannot deliver it.
    Monitor Direct Routing · Section 'Monitoring availability of Session Border Controllers using Session Initiation Protocol (SIP) options messages', demotion paragraphs · Checked 2026-09-25
  31. 31
    If 540998/503 occurs often, Microsoft says to check that the SBC re-resolves the Direct Routing FQDNs frequently, ideally every 15 minutes or more often.
    SIP 503 and Microsoft response codes - Microsoft Teams · Section '540998 503 Service Unavailable - Service state: Inactive/DrainingTransactions', second paragraph · Checked 2026-09-25
  32. 32
    The list of SBC response codes that trigger outbound failover is set per gateway with Set-CsOnlinePSTNGateway -FailoverResponseCodes, and any non-2xx final code can be configured.
    Trunk failover on outbound calls - Microsoft Teams · Section 'Failover of specific SIP codes received from the Session Border Controller (SBC)' · Checked 2026-09-25
  33. 33
    In Direct Routing, the default SBC final response codes that trigger retry of an outbound call on another SBC in the voice route are 408, 503 and 504.
    Trunk failover on outbound calls - Microsoft Teams · Section 'Failover of specific SIP codes received from the Session Border Controller (SBC)', last line · Checked 2026-09-25
  34. 34
    Direct Routing fails an outbound call over to another SBC only if the first SBC has sent no non-100 provisional response, to avoid double ringing.
    Trunk failover on outbound calls - Microsoft Teams · Section 'Failover of specific SIP codes received from the Session Border Controller (SBC)', first paragraph · Checked 2026-09-25
  35. 35
    When two or more SBCs in a Direct Routing route are healthy and of equal priority, calls are spread across them with a Fisher-Yates shuffle.
    Monitor Direct Routing · Section 'Monitoring availability of Session Border Controllers using Session Initiation Protocol (SIP) options messages', last paragraph · Checked 2026-09-25
  36. 36
    For Microsoft 365, Office 365 and GCC, the SBC should be configured with sip.pstnhub.microsoft.com, sip2.pstnhub.microsoft.com and sip3.pstnhub.microsoft.com in that priority order; GCC High and DoD each use a single FQDN.
    Plan Direct Routing · Sections 'SIP signaling: FQDNs', 'SIP signaling: GCC High', 'SIP signaling: DoD' · Checked 2026-09-25
  37. 37
    If Direct Routing cannot connect to a trunk (connection refused, TLS timeout or other network error), it retries the same trunk from a different Microsoft datacenter, which may be in another geographic region.
    Trunk failover on outbound calls - Microsoft Teams · Section 'Failover on network errors' · Checked 2026-09-25
  38. 38
    Direct Routing requires a successful OPTIONS exchange (200 OK) between the SBC and the Microsoft SIP proxy before calls can be set up.
    Teams Phone System Direct Routing: SIP protocol · Section 'Processing the incoming request: finding the tenant and user' · Checked 2026-09-25
  39. 39
    Direct Routing treats an SBC as healthy if it has sent SIP OPTIONS within the last three minutes, against an expected one-minute interval.
    Monitor Direct Routing · Section 'Monitoring availability of Session Border Controllers using Session Initiation Protocol (SIP) options messages' · Checked 2026-09-25
  40. 40
    A busy Direct Routing datacenter can send the SBC a Retry-After with a one-second interval.
    Teams Phone System Direct Routing: SIP protocol · Section 'Retry-After' · Checked 2026-09-25
  41. 41
    Teams Call Analytics only helps when a call reaches Direct Routing's internal components; for SBC pairing problems or rejected INVITEs, Microsoft points to the SBC logs, where Direct Routing sends detailed error descriptions.
    Monitor Direct Routing · Section 'Monitor Call Quality Analytics dashboard and SBC logs' · Checked 2026-09-25
  42. 42
    A Direct Routing outbound call that draws a 503 from the SBC after the SBC has already sent 180 or 183 will not fail over to another SBC even though 503 is a default failover code.inferred
    Trunk failover on outbound calls - Microsoft Teams · Section 'Failover of specific SIP codes received from the Session Border Controller (SBC)' · Checked 2026-09-25
  43. 43
    Because proxies are told not to relay 503 upstream and Teams reports whether a 503 came from Microsoft (540998) or the SBC (560503), the first troubleshooting step for a 503 is to find which hop generated it before blaming the far-end carrier.inferred
    RFC 5390: Requirements for Management of Overload in the Session Initiation Protocol · Synthesis of rfc-5390 section 4 and ms-learn-troubleshoot-dr-sip-503 sections 540998 and 560503 · Checked 2026-09-25
  44. 44
    Whether a 503 in reply to an OPTIONS keepalive takes a trunk down depends on the platform: Unified CM documents it as out of service, AudioCodes makes it configurable, and Teams Direct Routing health depends on whether the SBC keeps sending OPTIONS rather than on the response code.inferred
    Cisco Unified Communications Manager Trunks (Session Management Edition deployment guide 8.6(1)) · Synthesis of cisco-sme-861-trunks 'SIP OPTIONS Ping', audiocodes-sbc-um-740-proxy-sets 'Keep-Alive Failure Responses', ms-learn-dr-monitor SIP options section · Checked 2026-09-25
  45. 45
    RFC 3261 allows proxies to refuse the connection or drop the request instead of responding with 503.
    RFC 5390: Requirements for Management of Overload in the Session Initiation Protocol · Section 4 (quotation of RFC 3261) · Checked 2026-09-25
  46. 46
    RFC 3261 tells proxies not to forward 503 responses upstream, at SHOULD NOT strength.
    RFC 5390: Requirements for Management of Overload in the Session Initiation Protocol · Section 4 · Checked 2026-09-25
  47. 47
    After a 503, the client SHOULD NOT forward any other requests to that server for the duration given in the Retry-After header field, if one is present.
    RFC 5390: Requirements for Management of Overload in the Session Initiation Protocol · Section 4 (quotation of RFC 3261) · Checked 2026-09-25
  48. 48
    For server location, RFC 3263 defines failure as a 503 response, a transport failure (such as fatal ICMP errors on UDP or TCP connection failures), or the transaction timing out with no response.
    RFC 3263: Session Initiation Protocol (SIP): Locating SIP Servers · Section 4.3 · Checked 2026-09-25
  49. 49
    On such a failure the client SHOULD send an otherwise identical request with a new Via branch (a new transaction) to the next element in the DNS SRV-ordered list.
    RFC 3263: Session Initiation Protocol (SIP): Locating SIP Servers · Section 4.3 (also Section 4.2) · Checked 2026-09-25
  50. 50
    Under RFC 7339, an overloaded server that is throttling MUST reject some requests from non-participating clients with 503 without a Retry-After header.
    RFC 7339: Session Initiation Protocol (SIP) Overload Control · Section 5.10.2 · Checked 2026-09-25
  51. 51
    Under RFC 7339's default loss algorithm the oc value is a percentage from 0 to 100 by which the client must cut the requests it forwards to the overloaded server; oc-validity is in milliseconds.
    RFC 7339: Session Initiation Protocol (SIP) Overload Control · Sections 7.1 and 4.3 · Checked 2026-09-25
  52. 52
    RFC 7339 (Standards Track, September 2014) carries overload feedback hop by hop in Via header parameters oc, oc-algo, oc-validity and oc-seq; a client adds oc to the topmost Via it inserts and the server returns a value in the response's topmost Via.
    RFC 7339: Session Initiation Protocol (SIP) Overload Control · Sections 4.1 to 4.4 and 5.2 · Checked 2026-09-25
  53. 53
    Twilio Elastic SIP Trunking allows 1 call per second per trunk per region by default for termination; trunk-level CPS of 1 to 5 is self-service and above 5 requires Twilio sales.
    Calls per Second (CPS) — Trunking Termination · Default CPS and increase sections · Checked 2026-09-25
  54. 54
    When the Twilio termination CPS limit is exceeded, calls are rejected rather than queued, and the debugger logs error 32001 for trunk-level limits or 32012 for the pooled parent-account limit.
    Calls per Second (CPS) — Trunking Termination · CPS enforcement and debugger alerts sections · Checked 2026-09-25

Documents

tier 1 standards and regulators

RFC 3263: Session Initiation Protocol (SIP): Locating SIP Servers

RFC Editor / IETF · 2002-06-01 · accessed 2026-09-24

tier 1 standards and regulators

RFC 5390: Requirements for Management of Overload in the Session Initiation Protocol

IETF / RFC Editor · 2008-12-01 · accessed 2026-09-25

tier 1 standards and regulators

RFC 6357: Design Considerations for Session Initiation Protocol (SIP) Overload Control

IETF / RFC Editor · 2011-08-01 · accessed 2026-09-25

tier 1 standards and regulators

RFC 7339: Session Initiation Protocol (SIP) Overload Control

IETF / RFC Editor · 2014-09-01 · accessed 2026-09-25

tier 2 current vendor documentation

Calls per Second (CPS) — Trunking Termination

Twilio · 2026-07-21 · accessed 2026-09-24

tier 2 current vendor documentation

Calls through Session Initiation Protocol (SIP) Trunk Failure

Cisco · 2018-09-26 · accessed 2026-09-24

tier 2 current vendor documentation

Cisco Unified Border Element Configuration Guide - Cisco IOS XE 17.6 Onwards - SIP Trunk Monitoring

Cisco · 2026-04-25 · accessed 2026-09-24

tier 2 current vendor documentation

Cisco Unified Communications Manager Trunks (Session Management Edition deployment guide 8.6(1))

Cisco · 2010-12-17 · accessed 2026-09-24

tier 2 current vendor documentation

Configure Options Ping Between CUCM and CUBE

Cisco · 2019-02-06 · accessed 2026-09-24

tier 2 current vendor documentation

Configuring Proxy Sets - Mediant Software SBC User's Manual Version 7.4

AudioCodes · accessed 2026-09-25

tier 2 current vendor documentation

Monitor Direct Routing

Microsoft · 2026-09-10 · accessed 2026-09-21

tier 2 current vendor documentation

Plan Direct Routing

Microsoft (Microsoft Learn) · 2026-08-21 · accessed 2026-09-06

tier 2 current vendor documentation

SIP 503 and Microsoft response codes - Microsoft Teams

Microsoft · 2024-11-22 · accessed 2026-09-25

tier 2 current vendor documentation

Teams Phone System Direct Routing: SIP protocol

Microsoft · 2026-08-21 · accessed 2026-09-23

tier 2 current vendor documentation

Trunk failover on outbound calls - Microsoft Teams

Microsoft · 2025-05-06 · accessed 2026-09-25

Cite this page

APA

WarmTransfer. (2026, September 25). SIP 503 and trunk out-of-service failures. WarmTransfer. https://warmtransfer.net/knowledge/sip-503-trunk-failures

BibTeX

@misc{warmtransfer-sip-503-trunk-failures,
  title  = {SIP 503 and trunk out-of-service failures},
  author = {{WarmTransfer}},
  year   = {2026},
  url    = {https://warmtransfer.net/knowledge/sip-503-trunk-failures},
  note   = {Verified 2026-09-25}
}