Skip to content
PALM SOUQ

System status

Gateway uptime, region by region.

gateway.palmsouqltd.com runs on 280+ edge PoPs across five regions. SDKs fail open, so a gateway incident never blocks your provider calls.

All systems operational

Checked 08 Nov 2026, 14:00 UTC · probes every 30 s from 12 locations

Fleet uptime, 90d
99.9999%
Added latency p50
3.1 ms
Incidents, 90d
5

Regions

US East99.999% uptime
90 days agoToday

PoPs include iad, atl, mia, yyz

US West100.000% uptime
90 days agoToday

PoPs include sjc, sea, lax, den

EU Central100.000% uptime
90 days agoToday

PoPs include fra, ams, cdg, waw

India South100.000% uptime
90 days agoToday

PoPs include maa, blr, bom, hyd

Singapore100.000% uptime
90 days agoToday

PoPs include sin, kul, cgk, syd

No incidents Degraded performance Partial outage Maintenance

Components

  • Gateway proxyOpenAI-compatible and native provider endpointsOperational
  • Policy engine and breakersQuotas, rate limits, loop heuristicsOperational
  • Metering pipelineToken and cost attribution per tenantOperational
  • Billing syncStripe, Metronome, Orb and Lago exportsOperational
  • Webhooks and alertsSlack, PagerDuty and HTTPS endpointsOperational
  • Console and management APIDashboard, policy editor, REST APIOperational

Subscribe to incident updates by email or RSS from the console, or ask support@palmsouqltd.com to add your on-call address. Contractual SLAs are listed on the pricing page.

Incident history

  1. 2026-11-03 · All regions

    Scheduled maintenance: console database upgrade

    MaintenanceConsole and management API20 min
    1. In progress03 Nov 02:00 UTCMaintenance started. The console is read-only. Gateway traffic, breakers and metering are not affected.
    2. Completed03 Nov 02:20 UTCMaintenance completed. Policy edits are available again.

  2. 2026-10-21 · EU Central

    Elevated added latency in fra

    DegradedGateway proxy23 min
    1. Investigating21 Oct 14:06 UTCAdded latency at p50 rose to 9 ms for traffic terminating in fra. No requests failed.
    2. Identified21 Oct 14:15 UTCA noisy neighbour on one fra host. Traffic shifted to ams and cdg.
    3. Resolved21 Oct 14:29 UTCAdded latency back under 5 ms. Host replaced.

  3. 2026-09-30 · India South

    Delayed metering events in the console

    DegradedMetering pipeline41 min
    1. Investigating30 Sep 06:12 UTCPer-tenant spend in the console lagged by up to 6 minutes. Enforcement runs in the gateway and was not affected.
    2. Monitoring30 Sep 06:31 UTCConsumer group rebalanced. Backlog draining.
    3. Resolved30 Sep 06:53 UTCBacklog cleared. No metering events were lost.

  4. 2026-09-12 · US East

    Connection resets at one iad PoP

    Partial outageGateway proxy4 min
    1. Investigating12 Sep 19:40 UTCA share of connections to one iad PoP were reset. SDKs failed open to providers directly.
    2. Identified12 Sep 19:41 UTCFaulty load balancer drained. Anycast moved traffic to atl and yyz within 48 seconds.
    3. Resolved12 Sep 19:44 UTCAll PoPs healthy. Metering for fail-open calls was backfilled from SDK buffers.

  5. 2026-08-27 · Singapore

    Usage exports to Stripe delayed

    DegradedBilling sync58 min
    1. Identified27 Aug 03:10 UTCExports to Stripe Billing were retried with backoff after upstream rate limiting.
    2. Resolved27 Aug 04:08 UTCAll usage records delivered. No duplicate meter events.

Build on a gateway that fails open.

Every SDK falls back to the provider directly if the gateway is unreachable, and metering backfills when it returns. Your users never wait on PALM SOUQ.

14-day free production trial · 100k requests included · no card required

Start free trial Explore the demo