Uptime Monitoring Now in Open Beta 🚀🚀🚀

[gaprl]

We’re excited to announce Sentry's Uptime Monitoring is now available in open beta, enabling you effortlessly track the availability and reliability of your web services. Once configured, it continuously monitors your configured URLs, providing instant alerts and detailed insights to help you identify and resolve downtime promptly.

Set Up During Open Beta

Out of the box, Sentry automatically sets up Uptime Monitoring for the most frequently encountered hostname in your error data, ensuring continuous monitoring of your most critical services.

You can also create custom uptime monitoring alerts for specific URLs, with full control over request details like HTTP method, headers, and body parameters. Learn more how to set up custom uptime alerts or click here to create an alert.

How Uptime Monitoring Works

Sentry’s Uptime Monitoring performs health checks on configured URLs every minute. If an issue is detected, you'll be notified immediately, such as when:

• The URL returns a status code other than 200 (success).
• The response takes longer than 10 seconds (timeout).
• A DNS resolution error occurs.

By leveraging distributed tracing, Sentry enables you to pinpoint any errors that occur during an uptime check, streamlining triage and speeding up root cause identification. Learn more.

💬 We Want to Hear Your Feedback

If you have any thoughts or suggestions, please join the conversation below.👇

In case you're facing any issues, please open a new Github Issue.

[frank-nankivell]

Hi Team,

Really looking forward to using this feature on our platform - and ideally a set of Alerts we can the push to some external web-hooks. The feature appears to be removed from the UI. Aware this is in open beta so just checking on timelines for it returning?

Thanks!

[gaprl]

Hey @frank-nankivell, we had to temporarily disable the creation of new uptime alerts yesterday but we're back on. Feel free to click here to create a new uptime alert.

[gavinjohnson-crowdcube]

Are there any limitations or requisites to enable this? If I log into Sentry and navigate to https://[company].sentry.io/alerts/new/uptime , enter all the details and when I then hit the Create Alert button I get a 404 back from the POST to https://us.sentry.io/api/0/projects/[company]/[whatever project I'm in]/uptime/

[gaprl]

Hi @gavinjohnson-crowdcube, we had to temporarily disable the creation of new uptime alerts yesterday but we're back on. Feel free to click here to create a new uptime alert.

[gavinjohnson-crowdcube]

Yep, all working now!

[Anton-Aracor]

We have tow envs. But this alert is created automatically and sent notification to our PROD Slack channel, while pointing that DEV was downtime. In the configuration of the automatically created alert I see no option to change environment. And if I try to change url from http to https, for example, it din't save the change.

I need to have two separate alerts - one for PROD and one for DEV env. Didn't find how to do this configuration.

[gaprl]

Hey @Anton-Constructor, we had to temporarily disable the creation of new uptime alerts yesterday but we're back on. That's probably why you weren't able to save it. Can you please give it another try and let me know if you still face any issues?

[markmono]

Is there any simple way to exclude these events from the Discover events, as they spam it during my quiet times.

I do this for AWS ELB health checks by having a custom traces sampler in the python sentry_sdk checking for user_agent ELB-HealthChecker/2.0 and ignoring which I could also add SentryUptimeBot/1.0, but is there a better way to do this using a built in x-header which is always ignored by the sdk?

[gaprl]

Hey @markmono, yes using the User-Agent to filter uptime transactions in your SDK's tracer is the best way to go. Our SDKs do not have a built-in header that automatically ignores recording transactions.

[th3hamm0r]

For smaller plans, this does not seem to be an option, as these uptime requests seem to be counted 1:1 towards the transaction/performance-unit quota.
Are there any plans to change this, or to allow a higher interval (e.g. once per hour)?

[gaprl]

Hey @th3hamm0r, yes we'll be adding support for changing the interval.

Regarding the requests being counted towards your transaction quota, we recommend filtering on the client-side to not send any events from our uptime bot. You can use the requestUser-Agent to check if it's a Sentry bot. Do note that by not sending these transactions to Sentry, you won't be able to see its traces when an uptime issue is created.

[smo-qarma]

Looks quite interesting. As soon as you have any indicators for pricing, please let us know!

Also, do you plan to have uptime checks performed from different data centers around the world? We would be particularly interested in Asia and China...

[gaprl]

Hi @smo-qarma, we are considering offering the ability to run uptime checks in different regions around the world. If you can, I'd love to learn more about your use case. How would such feature help you and your team? What are some of the metrics and functionality you'd expect?

[mridang]

The feature is neat. Is there a way to configure the interval? It currently polls every second. I couldn't find such an option.

[gaprl]

Hey @mridang, we are shipping the ability to update the interval very soon. I'll post an update here as soon as it is live.

Now, you shouldn't be seeing checks running every second (they should be every minute). If that's the case, could you share with us the URL of your monitor? I can take a look into it. Feel free to email us at [email protected].

[gaprl]

Hey @mridang, just updating on my previous comment, the ability to update the interval is live and can be edited directly in your alert configuration.

[mridang]

Works like a charm

[smo-qarma]

Hi, thanks for getting back. We have a heavy presence in Asia, particularly China. Our goal is to make sure our availability in those regions is stable. So measuring uptime and also average response time. Today we use a simple service to do this which will perform a few retries if a location experiences an outage, and then we are alerted if it is a consistent outage. This is very basic. China is a story on its own in terms of access due to the Great Firewall, it is not straightforward to set up anything inside China. I will be happy to jump on a call if you need more info. We are very happy with Sentry and would happily help with input on moving more responsibility onto Sentry.

…

________________________________ From: Gabriel Lopes ***@***.***> Sent: Wednesday, October 2, 2024 20:39 To: getsentry/sentry ***@***.***> Cc: Søren Mønsted ***@***.***>; Mention ***@***.***> Subject: Re: [getsentry/sentry] Uptime Monitoring Now in Open Beta 🚀🚀🚀 (Discussion #78100) Hi @smo-qarma<https://github.com/smo-qarma>, we are considering offering the ability to run uptime checks in different regions around the world. If you can, I'd love to learn more about your use case. How would such feature help you and your team? What are some of the metrics and functionality you'd expect? — Reply to this email directly, view it on GitHub<#78100 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ASLCOPRMRHOQVDIXI564XSTZZQ4VPAVCNFSM6AAAAABOZNAPOWVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTAOBSGQZDMMQ>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[sandstrom]

I love Sentry, but the design of this feature seems really weird.

The main building block is an alert?

Generally how other uptime monitoring solutions work, is that you create the monitor with all it's details (URL, response checks, IPv6/IPv4, regions, etc) and you can then (optionally) tie alerts to those monitors.

But all uptime monitors doesn't have alerts tied to them, and also the criteria may differ. So e.g. you might want to run the monitor from 5 regions but only alert if its down from more than 3 at the same time.

It's also important with data insight, i.e. you need to be able to see for at least the past 24 hours or so what responses you are getting back.

I don't know if I've misunderstood how Sentry has designed it, but to me it looks like they haven't even looked at any of the myriad of existing tools for this (BetterUptime, Cronitor, Pingdom, Uptimerobot, etc).

[gaprl]

Hey @sandstrom, we really appreciate your feedback! We're taking a look how we can improve our UX, but rest assured some of the things you've mentioned are already on our roadmap, such as enhanced configurability for your monitors and history of checks.

[sandstrom]

Sounds good! I'll keep my eyes on the changelog.

Overall very happy with Sentry!

[rtshilston]

Can you provide instructions for how to see the performance over time of a given monitor? eg if I've set it to check api.mycompany.com/v1/health, how can I see the average response time / uptime / geographic split? I'd love to migrate from Pingdom to Sentry for this.

[Digital39999]

(public) Status pages soon? 👀