cloud fodder on Nostr: when i do servers professionally, if theres an outage i like to do a retrospective. ...
when i do servers professionally, if theres an outage i like to do a retrospective. so heres a nice short one for ya!
duration and impact: nostr1.com had approx 12 relays go down for 16 hours. including nostr21.com (one of the larger traffic relays). degraded response times also as, 1/3 of the haproxy layer was offline but not removed from round robin dns. (judgement call at the time as most web clients understand to try the next ip).
reason: contabo dropped a server and made me think it would come back #soon.
learned: how to better utilize contabo support, their status page lies, and they're still a "you get what you pay for" operation.
recommended changes: nothing, relays can't be expensive, big #cloud can still suck it, and pending further downtimes 16 hours of partial outage in 1.2 year span, worth. revisit on further downtime events.
duration and impact: nostr1.com had approx 12 relays go down for 16 hours. including nostr21.com (one of the larger traffic relays). degraded response times also as, 1/3 of the haproxy layer was offline but not removed from round robin dns. (judgement call at the time as most web clients understand to try the next ip).
reason: contabo dropped a server and made me think it would come back #soon.
learned: how to better utilize contabo support, their status page lies, and they're still a "you get what you pay for" operation.
recommended changes: nothing, relays can't be expensive, big #cloud can still suck it, and pending further downtimes 16 hours of partial outage in 1.2 year span, worth. revisit on further downtime events.