I would like updown to check every minutes, BUT notify me when it's down more than 20 minutes in a row, or, allow specific allowed down time, for backup reasons for example (once a week)
I need this feature as well. I need a way to measure and report on unplanned outages because it's part of our SLA. Is there any way to at least adjust the uptime stats to compensate for planned downtime?
I believe the declination of this idea should be reconsidered. For business users, it's important being able to report uptime outside of planned maintenance. Disabling monitoring manually for each maintenance window is not only impractical, it also removes the awareness and metrics offered by the monitoring plan. I want monitoring running during the maintenance, but to be able to flag a particular downtime interval after the fact to indicate it should not count towards uptime recorded. I came here looking for this particular feature and will have to keep looking, at least for now, despite what otherwise looks like an attractive service.
This was the biggest thing we missed when evaluating this product. Being able to publish this would drastically help add context to the outages and also be a great place to post them for companies that host their own web sites. I have attached a quick sketch of what I think would be most helpful in our use case. In the example, when the window is over it would automatically move to the bottom of the page. It would also be nice to manually add past windows for emergency maintenance that was not previously scheduled.
I'd like if i could add expected downtimes to a check.
Expected recurring downtimes: e.g. I check the internet connectivity of my customers and in Germany many internet service providers have a forced separation of the internet connectivity. It's always at a fix time, so the recurring downtime is plannable.
One-time: a way to just add an upcoming downtime. I know, this is possible with just pause the check, but not plannable.
Thanks!
Edit:
I forgot an important detail.
In the period of an expected downtime there just shouldn't come any notifications.
This is a simple example, you can easily extend this by only handling checks that weren't disabled before, or you can run this on set times using a cron job if you have a set maintenance window.
I like updown.io but not this feature is messing up our SLAs and causing alert fatigue so we've switching to an alternative, probably New Relic Synthetics.
I don't want to settle for scripting the API myself as there is a risk a monitor will be left disabled.
+1 it would be nice to have one-off maintenance (migrations, upgrades, ...) and repeating maintenance on a schedule (backups, credential rotations, ...) during which downtimes are ignored
Martijn We would like something similar too. For some servers we have a maintenance windows (e.g. last Sunday of every month from 9PM - 10PM) and it would be awesome if updown.io would be paused automatically during that time slot.
This would be similar to https://updown.canny.io/general/p/allow-maintance-window-for-planned-it-jobs. I guess people will need different things depending if they want to stop monitoring or just mute it, mute only the apdex or also the downtimes/SSL alerts/etc.. These are all features that would require a substancial addition to the UI and complexity only to serve a couple so it's unlikely to make it to the product as a new feature soon. But it's something I would encourage people to hack around using the API if they want this level of customisability out of updown.
In this example, one could use the API to tune down the Apdex alerts during the night (by increasing the Apdex thresold to 2 or 4 seconds for example) and restore the more sensitive setting in the morning. Similar to what people do to disable or mute checks at night.
Hi, Adrien Rey-Jarthon . I am not sure if it would be a feature request or a simple clarification. Currently we use API in a Github Action to mute alerts when maintenance window is turned on when deploying to Heroku. We use muting feature. This is all fine and dandy, but one thing that I noticed - looks like maintenance periods are counted towards downtime in our uptime metric.
Question: If I would switch to disabling the check instead, would that make sure that maintenance windows are not counted towards downtime?
Possible feature request: If not doable today, would be great to track two uptime metrics - the one that ignores maintenance (let's say the check was muted) and the one that does not. This way we could track deployment effect on the availability.
Crown Road Software thanks for reaching out, a small clarification first: if you use the API to mute alerts, as the name imply, the monitoring continues and count your downtimes as they occur (you just don't get alerts). It's not the "muted" time that counts as downtime though, only the actual downtimes.
If you switch to disabling the check instead, it would indeed stop the monitoring and thus "hide" the downtime, it will be considered as uptime. To be more precise: it'll be considered as uptime or downtime depending on the current state when you disable the check, so if you wait until the check is down before disabling, you're not gonna get uptime :)
And finally about the double uptime metric (including and excluding maintenance), that's unlikely to be added as it would be a bit too niche for updown.io design goals (simplicity & price). But with some work you can track those on your end using the downtime API, if you have deployments/maintenance kept on your end you can distinguish real downtimes and compute multiple uptime metrics however you want. I may add later a downtime comment feature though, so you can flag some downtimes as deployment or planned maintenance for your status page display and own recollection.
23 Comments
Put check in maintenance
I would like updown to check every minutes, BUT notify me when it's down more than 20 minutes in a row, or, allow specific allowed down time, for backup reasons for example (once a week)
Or to not monitor during certain times (like maintenance windows)
Great idea, it would be extra awesome if you can put in the periodical maintenance window.
I'm currently evaluating this product and this is a must have feature.
I need this feature as well. I need a way to measure and report on unplanned outages because it's part of our SLA. Is there any way to at least adjust the uptime stats to compensate for planned downtime?
I believe the declination of this idea should be reconsidered. For business users, it's important being able to report uptime outside of planned maintenance. Disabling monitoring manually for each maintenance window is not only impractical, it also removes the awareness and metrics offered by the monitoring plan. I want monitoring running during the maintenance, but to be able to flag a particular downtime interval after the fact to indicate it should not count towards uptime recorded. I came here looking for this particular feature and will have to keep looking, at least for now, despite what otherwise looks like an attractive service.
Ok I'm moving this back to "under review" as it's gathering a lot of interest, but this would require first https://updown.canny.io/general/p/log-of-uptime-downtime and https://updown.canny.io/general/p/allow-text-updates-for-incidents to be implemented to provide a decent UI for this.
An alternative way to doing this now is to automate the check disable/mute on your end using the API, quite easy to do in a maintenance script: https://updown.io/doc/how-to-automate-check-disable ( https://updown.io/doc/how-to-automate-check-disable )
This was the biggest thing we missed when evaluating this product. Being able to publish this would drastically help add context to the outages and also be a great place to post them for companies that host their own web sites. I have attached a quick sketch of what I think would be most helpful in our use case. In the example, when the window is over it would automatically move to the bottom of the page. It would also be nice to manually add past windows for emergency maintenance that was not previously scheduled.
@Adrien looks like second link there is admin-specific, but user-facing version appears to be https://updown.canny.io/general/p/allow-text-updates-for-incidents
that's a great idea! We really need this feature since we have scheduled maintenance times and now they're counted as "un-available" times
Hi,
first of all, i like it really.
I'd like if i could add expected downtimes to a check.
Expected recurring downtimes: e.g. I check the internet connectivity of my customers and in Germany many internet service providers have a forced separation of the internet connectivity. It's always at a fix time, so the recurring downtime is plannable.
One-time: a way to just add an upcoming downtime. I know, this is possible with just pause the check, but not plannable.
Thanks!
Edit:
I forgot an important detail.
In the period of an expected downtime there just shouldn't come any notifications.
For the meanwhile, you can easily fix this using the API ( https://updown.io/api ):
Get all your checks using https://updown.io/api/checks?api-key=:api_key
Loop through them and call
curl -X PUT -d 'enabled=false' https://updown.io/api/checks/:token?api-key=:api_key
to disable them.
This is a simple example, you can easily extend this by only handling checks that weren't disabled before, or you can run this on set times using a cron job if you have a set maintenance window.
I like updown.io but not this feature is messing up our SLAs and causing alert fatigue so we've switching to an alternative, probably New Relic Synthetics.
I don't want to settle for scripting the API myself as there is a risk a monitor will be left disabled.
This is already possible, both single-side and multi-site status page have the X-Frame-Options header remove to permit iframe embedding. With
+1 it would be nice to have one-off maintenance (migrations, upgrades, ...) and repeating maintenance on a schedule (backups, credential rotations, ...) during which downtimes are ignored
Already added comment in - https://updown.canny.io/general/p/allow-text-updates-for-incidents
This is a feature we would pay more money for.
One of our web apps reboots every single night at midnight, would be nice to schedule a mute window.
Martijn We would like something similar too. For some servers we have a maintenance windows (e.g. last Sunday of every month from 9PM - 10PM) and it would be awesome if updown.io would be paused automatically during that time slot.
This would be similar to https://updown.canny.io/general/p/allow-maintance-window-for-planned-it-jobs. I guess people will need different things depending if they want to stop monitoring or just mute it, mute only the apdex or also the downtimes/SSL alerts/etc.. These are all features that would require a substancial addition to the UI and complexity only to serve a couple so it's unlikely to make it to the product as a new feature soon. But it's something I would encourage people to hack around using the API if they want this level of customisability out of updown.
In this example, one could use the API to tune down the Apdex alerts during the night (by increasing the Apdex thresold to 2 or 4 seconds for example) and restore the more sensitive setting in the morning. Similar to what people do to disable or mute checks at night.
See: https://updown.io/doc/how-to-automate-check-disable
Hi, Adrien Rey-Jarthon . I am not sure if it would be a feature request or a simple clarification. Currently we use API in a Github Action to mute alerts when maintenance window is turned on when deploying to Heroku. We use muting feature. This is all fine and dandy, but one thing that I noticed - looks like maintenance periods are counted towards downtime in our uptime metric.
Question: If I would switch to disabling the check instead, would that make sure that maintenance windows are not counted towards downtime?
Possible feature request: If not doable today, would be great to track two uptime metrics - the one that ignores maintenance (let's say the check was muted) and the one that does not. This way we could track deployment effect on the availability.
Cheers, Edgars
Crown Road Software thanks for reaching out, a small clarification first: if you use the API to mute alerts, as the name imply, the monitoring continues and count your downtimes as they occur (you just don't get alerts). It's not the "muted" time that counts as downtime though, only the actual downtimes.
If you switch to disabling the check instead, it would indeed stop the monitoring and thus "hide" the downtime, it will be considered as uptime. To be more precise: it'll be considered as uptime or downtime depending on the current state when you disable the check, so if you wait until the check is down before disabling, you're not gonna get uptime :)
And finally about the double uptime metric (including and excluding maintenance), that's unlikely to be added as it would be a bit too niche for updown.io design goals (simplicity & price). But with some work you can track those on your end using the downtime API, if you have deployments/maintenance kept on your end you can distinguish real downtimes and compute multiple uptime metrics however you want. I may add later a downtime comment feature though, so you can flag some downtimes as deployment or planned maintenance for your status page display and own recollection.
Adrien Rey-Jarthon thanks for the reply, I will look into changing our integration to disable the checks instead of muting it.
Sign in to comment