Closed freefd closed 3 months ago
I've just faced the rare but painful side effect when the HTTP transport can do retransmission on the L3 but has no retransmissions for L4:
2024/06/23 17:40:35 +0000 INF ../tmp/src/notifier/telegram.go:69 > Alert sent event_id=1719153634.438606-w2iy47 provider=Telegram 2024/06/23 17:40:37 +0000 WRN ../tmp/src/notifier/nfty.go:64 > Unable to send alert error="Post \"https://ntfy.domain.tld/topicname\": dial tcp 192.168.xxx.xxx:443: connect: no route to host" event_id=1719153634.438606-w2iy47 provider=Ntfy
That happened due to reverse proxy pod in front of Ntfy was restarting and therefore unavailable.
Is it possible to add to both functions from https://github.com/0x2142/frigate-notify/blob/main/util/httpclient.go some reasonable retries, let's say 5 or even 6. So the maximum delay could be about 1 minute (6 retries * 10 seconds timeout).
Whether this needs to be configurable or not - idk.
Thank you.
Interesting - Thanks for reporting this. Should be easy enough to fix, I'll add it to the list!
Work like a charm, thanks a lot.
Awesome. Glad to hear it and thanks for the feedback!
I've just faced the rare but painful side effect when the HTTP transport can do retransmission on the L3 but has no retransmissions for L4:
That happened due to reverse proxy pod in front of Ntfy was restarting and therefore unavailable.
Is it possible to add to both functions from https://github.com/0x2142/frigate-notify/blob/main/util/httpclient.go some reasonable retries, let's say 5 or even 6. So the maximum delay could be about 1 minute (6 retries * 10 seconds timeout).
Whether this needs to be configurable or not - idk.
Thank you.