Re: Diagnosing outage
Also, its good to follow @herokustatus on twitter The new relic add-on has a ping feature which is useful. The StillAlive add-on goes a bit further with a full functional test (but doesn't run as frequently). On Fri, Jun 8, 2012 at 3:33 AM, Neil Middleton wrote: > You'll have trouble diagnosing this outage as the platform was at fault, > not your application. The routing layer was having issues for a few > minutes meaning that HTTP requests would not have been able to reach your > application (and as this happens in a layer above your application would > not show in your logs). > > Aside from this there are a number of logging options. The default Heroku > provided logging add-ons allow you to view up to 24 hours, whereas add-ons > such as Papertrail allow you to go back pretty much as far as you want (I > use Papertrail a fair amount on my apps and find it very useful) > > More info on Logging / Syslog (how Papertrail works) can be found here: > https://devcenter.heroku.com/articles/logging > > In the meantime, it's a good idea to bookmark the Heroku status site ( > http://status.heroku.com) as, generally speaking, any issues that are > outside of your control will appear here with more information. > > On Thursday, 7 June 2012 at 23:32, puzzler wrote: > > I've been testing out heroku, with a paid account, to determine > reliability. > Got a message from my pingdom account today that my server was down > for a half-hour. > > Looking through the logs for the time in question, I see nothing. > Specifically, I don't see the router process even reporting that it > received the pings for the half-hour in question. > > One possibly suspicious thing I see is that a half-hour before that, I > see a web dyno process exiting with code 143 and then restarting, and > I can't find any documentation about what that means. > > Another possibly suspicious thing is that in the two requests > immediately following the half-hour blank spot in the log, the > requests show a "wait time" of about 15ms. The amount seems > insignificant, but this web server sees so little traffic that the > wait time is always 0ms. I find it suspicious that the two requests > immediately following the outage would be the only requests I've ever > seen to show a wait time. > > The main thing I'm realizing is that I really have no idea how to > troubleshoot outages. The logs that show up when you type "heroku > logs" really don't go back that far. Is there a way to get them to go > back farther? How do I look up what specific exiting codes mean? Any > idea what could cause web requests to not get through to my app and > show up in the logs for a half-hour? > > Thanks, > > mark > > -- > You received this message because you are subscribed to the Google > Groups "Heroku" group. > > To unsubscribe from this group, send email to > heroku+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/heroku?hl=en_US?hl=en > > > -- > You received this message because you are subscribed to the Google > Groups "Heroku" group. > > To unsubscribe from this group, send email to > heroku+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/heroku?hl=en_US?hl=en > -- Thanks, -John -- You received this message because you are subscribed to the Google Groups "Heroku" group. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/heroku?hl=en_US?hl=en
Re: Diagnosing outage
You'll have trouble diagnosing this outage as the platform was at fault, not your application. The routing layer was having issues for a few minutes meaning that HTTP requests would not have been able to reach your application (and as this happens in a layer above your application would not show in your logs). Aside from this there are a number of logging options. The default Heroku provided logging add-ons allow you to view up to 24 hours, whereas add-ons such as Papertrail allow you to go back pretty much as far as you want (I use Papertrail a fair amount on my apps and find it very useful) More info on Logging / Syslog (how Papertrail works) can be found here: https://devcenter.heroku.com/articles/logging In the meantime, it's a good idea to bookmark the Heroku status site (http://status.heroku.com) as, generally speaking, any issues that are outside of your control will appear here with more information. On Thursday, 7 June 2012 at 23:32, puzzler wrote: > I've been testing out heroku, with a paid account, to determine > reliability. > Got a message from my pingdom account today that my server was down > for a half-hour. > > Looking through the logs for the time in question, I see nothing. > Specifically, I don't see the router process even reporting that it > received the pings for the half-hour in question. > > One possibly suspicious thing I see is that a half-hour before that, I > see a web dyno process exiting with code 143 and then restarting, and > I can't find any documentation about what that means. > > Another possibly suspicious thing is that in the two requests > immediately following the half-hour blank spot in the log, the > requests show a "wait time" of about 15ms. The amount seems > insignificant, but this web server sees so little traffic that the > wait time is always 0ms. I find it suspicious that the two requests > immediately following the outage would be the only requests I've ever > seen to show a wait time. > > The main thing I'm realizing is that I really have no idea how to > troubleshoot outages. The logs that show up when you type "heroku > logs" really don't go back that far. Is there a way to get them to go > back farther? How do I look up what specific exiting codes mean? Any > idea what could cause web requests to not get through to my app and > show up in the logs for a half-hour? > > Thanks, > > mark > > -- > You received this message because you are subscribed to the Google > Groups "Heroku" group. > > To unsubscribe from this group, send email to > heroku+unsubscr...@googlegroups.com > (mailto:heroku+unsubscr...@googlegroups.com) > For more options, visit this group at > http://groups.google.com/group/heroku?hl=en_US?hl=en > > -- You received this message because you are subscribed to the Google Groups "Heroku" group. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/heroku?hl=en_US?hl=en
Re: Diagnosing outage
Mark The loge tries addon goes back 24 hours. Jeff On Jun 7, 2012, at 5:32 PM, puzzler wrote: > I've been testing out heroku, with a paid account, to determine > reliability. > Got a message from my pingdom account today that my server was down > for a half-hour. > > Looking through the logs for the time in question, I see nothing. > Specifically, I don't see the router process even reporting that it > received the pings for the half-hour in question. > > One possibly suspicious thing I see is that a half-hour before that, I > see a web dyno process exiting with code 143 and then restarting, and > I can't find any documentation about what that means. > > Another possibly suspicious thing is that in the two requests > immediately following the half-hour blank spot in the log, the > requests show a "wait time" of about 15ms. The amount seems > insignificant, but this web server sees so little traffic that the > wait time is always 0ms. I find it suspicious that the two requests > immediately following the outage would be the only requests I've ever > seen to show a wait time. > > The main thing I'm realizing is that I really have no idea how to > troubleshoot outages. The logs that show up when you type "heroku > logs" really don't go back that far. Is there a way to get them to go > back farther? How do I look up what specific exiting codes mean? Any > idea what could cause web requests to not get through to my app and > show up in the logs for a half-hour? > > Thanks, > > mark > > -- > You received this message because you are subscribed to the Google > Groups "Heroku" group. > > To unsubscribe from this group, send email to > heroku+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/heroku?hl=en_US?hl=en -- You received this message because you are subscribed to the Google Groups "Heroku" group. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/heroku?hl=en_US?hl=en
Diagnosing outage
I've been testing out heroku, with a paid account, to determine reliability. Got a message from my pingdom account today that my server was down for a half-hour. Looking through the logs for the time in question, I see nothing. Specifically, I don't see the router process even reporting that it received the pings for the half-hour in question. One possibly suspicious thing I see is that a half-hour before that, I see a web dyno process exiting with code 143 and then restarting, and I can't find any documentation about what that means. Another possibly suspicious thing is that in the two requests immediately following the half-hour blank spot in the log, the requests show a "wait time" of about 15ms. The amount seems insignificant, but this web server sees so little traffic that the wait time is always 0ms. I find it suspicious that the two requests immediately following the outage would be the only requests I've ever seen to show a wait time. The main thing I'm realizing is that I really have no idea how to troubleshoot outages. The logs that show up when you type "heroku logs" really don't go back that far. Is there a way to get them to go back farther? How do I look up what specific exiting codes mean? Any idea what could cause web requests to not get through to my app and show up in the logs for a half-hour? Thanks, mark -- You received this message because you are subscribed to the Google Groups "Heroku" group. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/heroku?hl=en_US?hl=en