| commit | b7181a38c6b92bcb3b3cb17aa1f4c727987847dd | [log] [tgz] |
|---|---|---|
| author | Jan Kundrát <kundrat@cesnet.cz> | Thu Oct 31 12:40:23 2019 +0100 |
| committer | Jan Kundrát <kundrat@cesnet.cz> | Thu Oct 31 12:53:57 2019 +0100 |
| tree | f4d9e4c43a0ce2c5dcc5cc0b955a05b2fc59f7a7 | |
| parent | f253845e2ab27d2e06d50848697e370180a26e71 [diff] [blame] |
improve logging This allows for substantially more space for the log in an attempt to actually capture more useful data. Also ensure that whatever gets printed to stdout/stderr, such as C++'s own exception handling, is considered very important and dropped by accident. Also relax the rate limit a bit; at least that ITMA polling on the coherent A/D box produces a ton of data, and the trace bucket for that service gets throttled. A `journalctl -f -u systemd-journald` running in parallel was very useful in debugging this. Change-Id: If18446c9ab6c928a16e4944f331279cef627a612 Depends-on: https://cesnet-gerrit-czechlight/c/CzechLight/cla-sysrepo/+/1992 Depends-on: https://gerrit.cesnet.cz/c/CzechLight/cla-sysrepo/+/1992
diff --git a/package/cla-sysrepo/cla-appliance.service.in b/package/cla-sysrepo/cla-appliance.service.in index 7ee5c31..c48bce3 100644 --- a/package/cla-sysrepo/cla-appliance.service.in +++ b/package/cla-sysrepo/cla-appliance.service.in
@@ -18,6 +18,9 @@ SystemCallArchitectures=native Restart=always WatchdogSec=15s +LogRateLimitIntervalSec=10 +LogRateLimitBurst=30000 +SyslogLevel=alert [Install] WantedBy=multi-user.target