Open joshk opened 16 hours ago
@whatyouhide I think this should just be wrapped in a try/catch statement. Yes?
def attach_telemetry_handler(config) when is_list(config) do
_ = :telemetry.attach_many(__MODULE__, @events, &__MODULE__.handle_event/4, config)
:ok
end
@savhappy definitely not, that's not what the stacktrace is pointing to. :telemetry.attach_many/4
doesn't fail, it can just return :ok | {:error, ...}
so we don't really care about that.
Error comes from the stacktrace in the issue:
Stacktrace=[\n {Sentry.Client, :maybe_log_send_result, 2,\n [file: ~c\\"lib/sentry/client.ex\\", line: 370]} ...
Environment
Version 10.7.1 Oban cron integration enabled
Steps to Reproduce
The monitoring of Oban crons work perfectly, but when there is a Sentry outage, we have found that crons continue not to report even once the outage is resolved.
After looking into our logs, we found this:
The message above was after a bunch of
Failed to send Sentry event. Received 503 from Sentry server
messages.Once our app is restarted everything works as expected and the crons are reported as healthy.
Expected Result
The SDK should recover from Sentry outages.