osmocom
/
libosmocore
mirror of https://gerrit.osmocom.org/libosmocore
14
0
Fork 1
Code Issues Releases Wiki Activity

refactor stat_item: report only changed values 

Change the functionality of skipping unchanged values: instead of
looking up whether new values have been set on a stat item, rather
remember the last reported value and skip reporting identical values.

stats_test.c shows that previously, a stat item reported a value of 10
again, even though the previous report had already sent a value of 10.
That's just because the value 10 was explicitly set again, internally.

From a perspective of preserving all data points, it could make sense to
send consecutive identical values. But since we already collapse all
data points per reporting period into a max, that is pointless.

Related: SYS#5542
Change-Id: I8f4cf34dfed17e0879716fa2cbeee137c158978b
This commit is contained in:
Neels Hofmeyr 2021-09-14 21:49:00 +02:00 committed by neels
parent e90c7176be
commit 6a5940740a
4 changed files with 9 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
TODO-RELEASE
View File

@ -20,3 +20,4 @@ libosmocore osmo_stat_item API breakage: remove functions osmo_stat_item_get_ne
libosmocore osmo_stat_item_value API breakage: struct definition removed, because no known users exist / no users should exist.
libosmocore osmo_stat_item ABI breakage: struct osmo_stat_item made opaque.
libosmocore osmo_stat_item No FIFO buffer of values used anymore, the "skipped values" error is no longer possible.
libosmocore stats reporting for osmo_stat_item, values are now never repeated from one stats report to the next.

12
src/stats.c
View File

@ -700,15 +700,11 @@ static int osmo_stat_item_handler(
if (!srep->running)
continue;
/* If no new stat values have been set in the current reporting period, skip resending the value.
* However, if the previously sent value is not the same as the current value, then do send the changed
* value (while n == 0, the last value from the previous reporting period is in item->value.max == .last
* == .min, which was put in new_value above).
* Also if the stats reporter is set to resend all values, also do resend the current value regardless
* of repetitions.
/* If the previously reported value is the same as the current value, skip resending the value.
* However, if the stats reporter is set to resend all values, do resend the current value regardless of
* repetitions.
*/
if ((!item->value.n && new_value == prev_reported_value)
&& !srep->force_single_flush)
if (new_value == prev_reported_value && !srep->force_single_flush)
continue;
if (!osmo_stats_reporter_check_config(srep,

4
tests/stats/stats_test.c
View File

@ -443,10 +443,10 @@ static void test_reporting()
osmo_stat_item_set(osmo_stat_item_group_get_item(statg1, TEST_A_ITEM), 10);
do_report(0, 2);
fprintf(stderr, "report (group 1, item 1 update twice):\n");
fprintf(stderr, "report (group 1, item 1 update twice, with same value):\n");
osmo_stat_item_set(osmo_stat_item_group_get_item(statg1, TEST_A_ITEM), 10);
osmo_stat_item_set(osmo_stat_item_group_get_item(statg1, TEST_A_ITEM), 10);
do_report(0, 2);
do_report(0, 0);
fprintf(stderr, "report (group 1, item 1 update twice, check max):\n");
osmo_stat_item_set(osmo_stat_item_group_get_item(statg1, TEST_A_ITEM), 20);

6
tests/stats/stats_test.err
View File

@ -114,10 +114,8 @@ report (group 1, item 1 update):
test2: item p= g=test.one i=1 n=item.a v=10 u=ma
test1: item p= g=test.one i=1 n=item.a v=10 u=ma
reported: 0 counter vals, 2 stat item vals
report (group 1, item 1 update twice):
test2: item p= g=test.one i=1 n=item.a v=10 u=ma
test1: item p= g=test.one i=1 n=item.a v=10 u=ma
reported: 0 counter vals, 2 stat item vals
report (group 1, item 1 update twice, with same value):
reported: 0 counter vals, 0 stat item vals
report (group 1, item 1 update twice, check max):
test2: item p= g=test.one i=1 n=item.a v=20 u=ma
test1: item p= g=test.one i=1 n=item.a v=20 u=ma