Overall uptime
90.403%
last 90 days
Active incidents
0
all clear
Services tracked
8
public + private
Mean time to resolve
24 m
rolling 90 days
// services
POST /v1/email and the synchronous send pipeline.
Outbound MTA pool — dedicated + shared IPs.
Outbound HTTP delivery + replay queue.
app.voltmail.io — analytics + admin UI.
Four bare-metal nodes across us-east-1 + eu-west-1.
Customer + suppression + log metadata stores.
Idempotency cache + job queues.
Marketing site + dashboard JS/CSS.
// recent incidents
Every incident gets a public phase-by-phase write-up. Post-mortems are linked from the Resolved phase whenever published.
Investigating · 07:23 UTC
Synthetic checks observed elevated 5xx on the EU send pipeline.
Identified · 07:31 UTC
Root cause: a stuck Postgres replica drifting on idempotency-cache writes. Failing over to the secondary.
Monitoring · 07:42 UTC
Failover complete. Error rate back to baseline.
Resolved · 07:51 UTC
Replica recovered, monitored for 9 minutes, declaring resolved. Post-mortem: voltmail.io/post-mortems/2026-04-15.
Investigating · 14:08 UTC
p95 webhook delivery latency rose above 8s. Investigating queue depth.
Identified · 14:21 UTC
A noisy customer with 14k retried-out endpoints saturated the BullMQ DLQ workers.
Monitoring · 14:30 UTC
Per-tenant queue isolation rolled forward. Latency back to <1s.
Resolved · 14:42 UTC
No customer impact lasted >5 minutes; SLA credits issued automatically.
Maintenance · 02:00 UTC
Pre-announced read-only window for Postgres 16 → 17 upgrade.
Resolved · 02:23 UTC
Upgrade complete. Read+write restored. No data loss.
Investigating · 18:14 UTC
Some users see "ChunkLoadError" on app.voltmail.io.
Identified · 18:17 UTC
CDN cache mismatch after deploy — old chunk hashes pointed to purged objects.
Resolved · 18:21 UTC
Cache fully purged + rolling deploy of fresh hashes. Issue resolved.