[07:50:01] hey :) [07:51:02] hop [07:53:44] so, to fill in the missing pieces: internally ATS has no distinction between miss and pass. We could set 'pass' ourselves when we force the no-cache decision in lua, but if the response isn't cached by ATS due to no-cache response headers, that's still gonna show up as a miss [07:56:47] and cdanis is right, all traffic including misc services goes through ATS [09:19:43] <_joe_> heads up: I'm adding a new service to LVS [09:19:58] <_joe_> I will restart pybal low-traffic in codfw and eqiad in the next hour or so [09:20:50] ack [09:35:34] <_joe_> I was thinking [09:35:53] <_joe_> would it make sense to have pybal expose the status of its bgp advertisements via the http api [09:36:11] <_joe_> given that would allow us to automate restarts a bit better? [09:37:11] maybe, yeah [09:37:23] bgp status is exposed via prometheus metrics right now [09:37:28] <_joe_> oh [09:37:34] <_joe_> that's good enough :) [09:37:45] i've implemented that in what it feels another life already [09:38:20] :) [09:38:25] <_joe_> pybal_bgp_session_established{local_asn="64600",peer="208.80.153.192"} 1.0 [09:38:28] <_joe_> wonderful [09:38:31] <_joe_> thanks [09:38:32] that's it :) [09:39:01] <_joe_> ok, cookbook for pybal restarts coming up :D [14:45:49] 10Acme-chief, 10Traffic, 10Operations: Memory leak on acme-chief 0.21 - https://phabricator.wikimedia.org/T234131 (10MoritzMuehlenhoff) @Vgutierrez I created a 2.6.1-3+deb10u2, it's in my home on acmechief1001. Let's deploy this on acmechief* hosts on Monday before I submit this to the Debian stable release... [15:13:14] 10netops, 10Operations: IRR updates needed - https://phabricator.wikimedia.org/T235886 (10ayounsi) p:05Triage→03Low [18:14:19] 10Traffic, 10Operations, 10ops-ulsfo: ps1-22-ulsfo & ps1-23-ulsfo - https://phabricator.wikimedia.org/T235911 (10RobH) p:05Triage→03Normal