Attention is currently required from: iedemam, neels. pespin has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/27081 )
Change subject: WIP: New stats for lchan life duration. ......................................................................
Patch Set 4:
(5 comments)
File src/osmo-bsc/bts_trx_vty.c:
https://gerrit.osmocom.org/c/osmo-bsc/+/27081/comment/7ecf102a_5a8ad6b4 PS4, Line 575: vty_out(vty, " Activated %llu ms ago%s", duration_ms, VTY_NEWLINE); Probably more interesting to print it in VTY as seconds? Doesn't look like we can glance millisecond updates in the VTY anyway.
File src/osmo-bsc/bts_vty.c:
https://gerrit.osmocom.org/c/osmo-bsc/+/27081/comment/0fd3e91a_42ca9366 PS4, Line 3830: vty_out(vty, " Channel Activations : %"PRIu64" TCH", activations_tch); I'd rather put it: Channel Activations: TCH (avg lifespan 3s) SDCCH (avg lifespan 5s)
See the line formatting, and also use seconds instead of milliseconds. "Average" can probably be shortened to "avg".
File src/osmo-bsc/lchan_fsm.c:
https://gerrit.osmocom.org/c/osmo-bsc/+/27081/comment/2f4a8537_0e73101b PS4, Line 1812: long long duration_ms = gsm_lchan_active_duration_ms(lchan); Shouldn't this be an unsigned long long everywhere? Maybe simply use time_t?
https://gerrit.osmocom.org/c/osmo-bsc/+/27081/comment/0722481b_de0b326b PS4, Line 1814: if (lchan->type == GSM_LCHAN_TCH_F || lchan->type == GSM_LCHAN_TCH_H) { I'd go for a switch statement here, but not critical.
https://gerrit.osmocom.org/c/osmo-bsc/+/27081/comment/f9e99713_697ad89f PS4, Line 1816: rate_ctr_add(rate_ctr_group_get_ctr(bts_ctrs, BTS_CTR_LCHAN_TCH_TOTAL_ACTIVE_MILLISECONDS), duration_ms); Does it really makes sense to use this kind of value as a rate_ctr? Just asking. It's difficult to interpet it I think.