3 ms·
A lovely thing about tail latency in the chains that the post talks about is how one service being slow can cascade. Especially in serial chains, when on compon
by cratermoon 6y ago
A lovely thing about tail latency in the chains that the post talks about is how one service being slow can cascade. Especially in serial chains, when on component is slow, the rest are waiting, in the meantime using resources like memory, sockets, cpu cycles, that could be used to service other requests. In the worst cases, those other services can start responding slowly to other requests, resulting in further degradation.
Having circuit breakers and carefully tuning timeouts can help.
- mjb 6y ago> Having circuit breakers and carefully tuning timeouts can help. Can help, for sure. Can also turn a 'slow' into 'unavailable', which could be what you want, or could be a disaster.
- cratermoon 6y agoI'd take an 'unavailable' lasting for a short time while connections exponentially back off with jitter over minutes or hours of degraded service for everyone.