Commit graph

153 commits

Author SHA1 Message Date
Brian Gough
e5747fefd2 remove package-lock.json (not needed for library) 2020-08-06 14:52:07 +01:00
Brian Gough
5b39d49358 use scoped package name @overleaf/metrics 2020-08-05 11:27:56 +01:00
Brian Gough
d415ae0cbe remove gruntfile 2020-07-17 16:23:01 +01:00
Brian Gough
a0f856cff2 fix tests 2020-07-17 16:17:18 +01:00
Brian Gough
747a80b545 decaffeinate 2020-07-17 16:01:58 +01:00
Brian Gough
e31a819636 remove statsd 2020-07-17 15:36:37 +01:00
Eric Mc Sween
e197e4dc11 2.7.0 2020-07-06 08:20:02 -04:00
Eric Mc Sween
d69195eaa9 Log requests that don't have a route property
The v1 history service has its routes set up via swagger-tools, which
doesn't write a route property on the request. This prevents us to send
request metrics based on the route, but we can still log the request.
2020-07-03 16:38:29 -04:00
Jakob Ackermann
ea767920fc [misc] fix unit tests 2020-04-01 16:40:15 +02:00
Eric Mc Sween
8131c5ac91 2.6.2 2020-03-25 09:53:54 -04:00
Eric Mc Sween
ef7e3b0e7a Fix HTTP request timing metrics
The conversion between high resolution time and milliseconds was
incorrect.
2020-03-25 09:39:21 -04:00
Eric Mc Sween
a17843f3bf 2.6.1 2020-03-20 07:47:05 -04:00
Eric Mc Sween
25448bfef4 Downgrade all request logs to INFO
Commit 9056143fe36f1347a1ff985ef8592a1de7d798dd added logic to log
requests with different error levels depending on the status code. The
intention was to make the 5xx and 4xx requests stand out in Stackdriver.
Unfortunately, this also creates a lot of noise in Sentry since we log
the errors separately from the requests.

This commit brings back the former behaviour of logging all requests at
the INFO level. We can revisit this if we integrate the strategy
implemented in filestore of logging once per request.
2020-03-18 08:41:11 -04:00
Eric Mc Sween
62e82d7469 Send the status code with the HTTP request size metric 2020-03-12 06:24:46 -04:00
Eric Mc Sween
4a92be80ea Send HTTP request size metric
The metric is a "summary" called http_request_size_bytes.
2020-03-11 16:39:49 -04:00
Brian Gough
03e81153db avoid step effects in summary metrics
reduce the window size from 10 minutes to 1 minute,  so that short
spikes do not cause a 10 minute long "table" graph.
2020-03-10 15:01:09 +00:00
Simon Detheridge
74c2afc12d Bump package version 2020-03-04 10:31:12 +00:00
Simon Detheridge
9a8cddbbb6 Don't set UV_THREADPOOL_SIZE if already set 2020-03-03 17:09:35 +00:00
Brian Gough
2273978e7b fix gauge usage 2019-12-16 11:42:10 +00:00
Brian Gough
3a5374c6f9 increase minor version for backwards compatible addition 2019-12-16 10:22:50 +00:00
Brian Gough
93df87eff3 allow options for count 2019-12-16 10:18:30 +00:00
Simon Detheridge
53aa2490f5 Merge pull request #20 from overleaf/spd-metrics-ttl
Add mechanism to expire old prometheus metrics
2019-10-28 17:21:57 +00:00
Simon Detheridge
feecda8ea8 Use map instead of hash for metrics 2019-10-28 14:39:53 +00:00
Simon Detheridge
e0cf10a886 Fix typo, gague -> gauge 2019-10-28 12:34:04 +00:00
Eric Mc Sween
7b7b6d0793 2.3.0 2019-10-25 07:17:43 -04:00
Eric Mc Sween
166211b278 Stackdriver logging
When the environment variable STACKDRIVER_LOGGING is set to true, send
request logs in a format that Stackdriver knows how to interpret. Also,
set the log level accordint to the status code. 4xx responses are logged
as warnings and 5xx responses are logged as errors.
2019-10-24 22:05:12 -04:00
Simon Detheridge
07e4eb4dfb Add mechanism to expire old prometheus metrics
Adds a wrapper around the prometheus client, which keeps track of the
last time a metrics was accessed, and removes old ones once they have
not been accessed for a period of time.
2019-10-23 17:07:45 +01:00
Brian Gough
286eb747ad add status label to gauges 2019-06-06 10:34:55 +01:00
Henry Oswald
816c49daf2 bump metrics to 2.1.2 2019-04-12 12:36:47 +01:00
Henry Oswald
d4faaaa60e use console.log not logger.log
we don't habe verbiouse lossing enabled that early in the apps lifecycle
also best not to require modules before enabling profiler
2019-04-12 12:34:17 +01:00
Henry Oswald
15d14d8e2b add injectMetricsRoute into statsd so it doens't blow up 2019-02-07 09:47:29 +00:00
Henry Oswald
bf18c6e513 MVP for running both statsd and prom side by side
statsd code is from v1.8.1
2019-01-30 11:11:37 +00:00
Henry Oswald
85011ed0e7 add path into inc options 2019-01-28 14:37:54 +00:00
Christopher Hoskin
02907fd2e7 Fix Register II 2019-01-15 16:16:12 +00:00
Christopher Hoskin
20e45b7a2e Fix Register 2019-01-15 16:12:15 +00:00
Christopher Hoskin
d4caa48118 Bump package version 2019-01-15 15:44:01 +00:00
Christopher Hoskin
93bef54c39 Make register public so we can support other servers e.g. hapi 2019-01-15 15:36:35 +00:00
Henry Oswald
dfead32d69 Update package.json 2018-12-13 09:19:05 +00:00
Henry Oswald
1a34f3db1b Update package.json 2018-12-13 08:46:58 +00:00
Henry Oswald
287effb139 use ENABLE_PROFILE_AGENT 2018-12-12 21:17:12 +00:00
Henry Oswald
db4ae84bda require logger 2018-12-12 21:05:33 +00:00
Henry Oswald
a86c4d8abd add profiler 2018-12-12 20:11:40 +00:00
Henry Oswald
9eac49ad84 add some logging in 2018-12-11 16:07:34 +00:00
Henry Oswald
c257482e15 add ENABLE_DEBUG_AGENT and don't require modules unless they are enabled 2018-12-11 15:46:29 +00:00
Christopher Hoskin
6f82309829 Bump version to 2.0.10 2018-12-11 12:52:38 +00:00
Christopher Hoskin
61e6cf0493 Add host label to timing metrics 2018-12-11 12:01:22 +00:00
Henry Oswald
425a6f55ff set build version via env var and bump build version 2018-12-10 22:02:25 +00:00
Henry Oswald
7a227adaeb add if statment around traceagenet and bump to 2.0.8 2018-12-05 13:58:40 +00:00
Henry Oswald
02071584ae bump to 2.0.7 2018-12-05 12:31:16 +00:00
Henry Oswald
b2f49351c0 few tidy up changes 2018-12-05 11:03:40 +00:00