Patikimumas ir eksploatacija
HansaChat sistemos būsena.
Tiesioginis prieinamumas, išmatuotas veikimo laikas ir visi užfiksuoti incidentai – tikrinama nepriklausomu išoriniu monitoriumi kas penkias minutes.
- Pokalbių paslaugos veikimo laikas, paskutinės 30 dienų
- 99.89%
- Pokalbių paslaugos veikimo laikas, paskutinės 90 dienų
- 99.95%
- Prieinamumo tikslas
- 99,9 % kas mėnesį
- Stebėjimas
- Nepriklausomas, kas 5 minutes
Dabartinė būsena
Kiekviena pagrindinė paslauga, tikrinama kas penkias minutes.
Monitorius kreipiasi į pokalbių būklės (health) gaunamąjį ir viešą svetainę tiesiogiai ir atskirai užfiksuoja kiekvienos pagrindinės paslaugos rezultatą.
Paskutinis patikrinimas: 08:25 UTC · 441 ms atsako laikas
- Pokalbių API Veikia
- Duomenų bazė Veikia
- Podėlis Veikia
- Žinučių eilė Veikia
- Realtime Veikia
- Svetainė Veikia
Išmatuotas veikimo laikas
Devyniasdešimt dienų matavimų, diena po dienos.
Kiekviena juosta yra viena diena. Prieinamumas yra sėkmingų penkių minučių patikrinų dalis – matuojama išorinio monitoriaus, ne pačios HansaChat. Paslaugų lygio sutartis vertina prieinamumo tikslą kalendoriniam mėnesiui; kasdienės juostos rodo slankiuosius 30 ir 90 dienų langus nuolatiniam matomumui.
Pokalbių paslauga
30 dienų: 99.89% 90 dienų: 99.95%
8,422 patikrinimų per 30 dienų
Svetainė
30 dienų: 99.91% 90 dienų: 99.95%
Incidentų istorija
Visi užfiksuoti incidentai, su post-mortem analizėmis.
Incidentai atveriami automatiškai po trijų iš eilės nepavykusių patikrinų ir uždaromi, kai kitas patikrinimas sėkmingas. Reikšmingi incidentai gauna paskelbtą post-mortem analizę.
-
Išspręstas Svetainė rgs 29, 2026, 08:00 UTC · 24 min. 50 sek.
HansaChat Status Alert Landing Page has been DOWN for 3 consecutive checks. Affected services: Landing Page Last error: HTTP 504 Time: 2026-09-29T08:00:39.412Z
Post-mortem
Summary On 2026-09-29 both the chat API and the landing page were unavailable for about 30 minutes. Automated 5-minute checks recorded continuous HTTP 502/504 gateway failures between 07:50 and 08:25 UTC. Root cause An IONOS Cloud incident: partial connectivity degradation to the control plane affecting Kubernetes operations. See https://status.ionos.cloud/incidents/zprjxl2b1kmy for details. No defect in HansaChat application code was involved. Impact Around 30 minutes of downtime for both the chat API and the landing page; all incoming requests failed with HTTP 502/504 gateway errors. Resolution Service recovered automatically once the IONOS incident ended. Monitoring confirmed full recovery at 08:25 UTC; no HansaChat-side intervention was required.
-
Išspręstas Pokalbių paslauga rgs 29, 2026, 08:00 UTC · 24 min. 49 sek.
HansaChat Status Alert Chat service has been DOWN for 3 consecutive checks. Affected services: Chat API Last error: HTTP 504 Time: 2026-09-29T08:00:38.344Z
Post-mortem
Summary On 2026-09-29 both the chat API and the landing page were unavailable for about 30 minutes. Automated 5-minute checks recorded continuous HTTP 502/504 gateway failures between 07:50 and 08:25 UTC. Root cause An IONOS Cloud incident: partial connectivity degradation to the control plane affecting Kubernetes operations. See https://status.ionos.cloud/incidents/zprjxl2b1kmy for details. No defect in HansaChat application code was involved. Impact Around 30 minutes of downtime for both the chat API and the landing page; all incoming requests failed with HTTP 502/504 gateway errors. Resolution Service recovered automatically once the IONOS incident ended. Monitoring confirmed full recovery at 08:25 UTC; no HansaChat-side intervention was required.
-
Išspręstas Pokalbių paslauga rgs 2, 2026, 09:15 UTC · 5 min.
HansaChat Status Alert Chat service has been DOWN for 3 consecutive checks. Affected services: Chat API Last error: HTTP 502 Time: 2026-09-02T09:15:53.599Z
Post-mortem
Summary The real-time chat API was unavailable for about 15 minutes on 2026-09-02 (09:05-09:20 UTC). Requests to the chat API returned HTTP 502 because the queue processor worker could not start. The landing page stayed reachable apart from a single failed probe during the same window. Root cause HansaChat started as a startup, and the initial credentials-injection implementation fetched credentials from the upstream API on every request, without any caching. During an unusually intense development day more than 1,000 items were requested in a short window, which exhausted the upstream API's rate limit. Without credentials available the queue processor worker could not start, which took the real-time chat API down with it. Impact Roughly 15 minutes of real-time chat API downtime (09:05-09:20 UTC); queued message processing was stalled for the same window. The landing page was not materially affected. Resolution Service recovered once the upstream rate limit window reset (09:20 UTC). Credentials injection now uses an in-cluster cache, so a temporary upstream rate limit can no longer prevent worker startup.
-
Išspręstas Svetainė rgp 14, 2026, 18:45 UTC · 9 min. 45 sek.
HansaChat Status Alert Landing Page has been DOWN for 4 consecutive checks. Affected services: Landing Page Last error: The operation was aborted due to timeout Time: 2026-08-14T18:45:41.129Z
Post-mortem
Summary The landing page was unavailable for about 20 minutes on 2026-08-14 (18:35-18:55 UTC). Chat services recorded no downtime during the incident. Root cause The underlying Kubernetes node failed and its pods were rescheduled. Traefik kept routing correctly to the old pods, which were still serving while being drained. After the deployment of the new landing version, Traefik could not find the new pods' IP addresses, so landing traffic was sent to endpoints that no longer existed. Impact Landing page downtime of roughly 20 minutes. No chat downtime was recorded.
Nepriklausomas monitorius
Sukurtas taip, kad toliau fiksuotų, kai neveikia pati HansaChat.
Šis būsenos puslapis talpinamas kartu su produktu. Matavimus renka atskiras Cloudflare worker, nepriklausomas nuo HansaChat DNS, talpinimo regiono ir infrastruktūros – jei hansa.chat kada nors taptų nepasiekiama, monitorius toliau tikrina, o jo puslapis toliau atsako.