Slack Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

Slack Issues Reports

Latest outage, problems and issue reports in social media:

• 
  Sophia Yen MD (@teenmd) reported 3 minutes ago

  .@SlackHQ Why don't you have human service agents? You are a GIANT company used by many. The only contact us is via a webform?

• 
  Ash (@edweirdsnowman) reported 4 minutes ago

  @SlackHQ slack huddles has had video audio sync issues since it’s been launched. Please deploy at least 7 agents to fix this immediately.

• 
  dnd dhmu (@refreshmygrave) reported 6 minutes ago

  @advaithj1 @thdxr @SlackHQ bro pls fix ur support for discord no real people reply ever

• 
  Vortex (@mjparme) reported 8 minutes ago

  @SlackHQ why is there no obvious way to join a channel in Slack? Seems like a huge usability problem, the + sign next to channels creates a channel, but joining one is very elusive

• 
  Cass Anderson (@casspa) reported 10 minutes ago

  @Yonar87 @SlackHQ same issue here. Just sent them an email. Their status landing page says they've been 100% find this quarter which makes me think maybe they just don't update that page?

• 
  Aadii (@AadityaPasayat) reported 12 minutes ago

  @raidingAI @SlackHQ @slashyai I am sorry to say, but u don't have taste, either the editor. R eyes go from left to right, common sense. Btw I can help u get taste

• 
  Chris Munns (@chrismunns) reported 14 minutes ago

  somehow @SlackHQ filtering of channels/DMs for last 30 days active hasn't worked for me for months. is this just a broken feature or have I done a thing.

• 
  Jon Slemp (@Jon_Slemp) reported 16 minutes ago

  Bug report for you @SlackHQ — when receiving a DM notification from Agents, and those agents are setup as Slack apps you have to DM directly in an app channel, once you hit a certain number of chats in your history tab you can no longer navigate to the DM directly by clicking the DM notification on mobile. You basically get navigated to a blank screen and effectively chat find the message ever. Please fix. This is a disaster of a bug.

• 
  jimmy hackett (@j1mmyhackett) reported 19 minutes ago

  i shipped a report that posted a PDF to @SlackHQ with buttons to approve or flag it. worked great in testing. turns out every single report was re-sharing the exact same file. Next.js had cached the upload URL on the first request and just kept reusing it. so every property got the same PDF. same charts. same data. looked fine in the message. was completely wrong. the fix was straightforward once i saw it. but i only saw it because i added logging on the upload path and watched what Slack was actually returning. without that i would have kept assuming the reports were right. this is the thing that keeps biting me. i build something, it posts, the message looks correct, and i ship it. but "the message looks correct" is not the same as "the data is correct". i need the system to surface the difference, not just tell me a message was sent. the reports now work the way they were supposed to. one message per property, each with its own PDF, two models reading it and leaving grades in the thread. the buttons are how you reply to the report, not how you open it. but i almost shipped a monitoring system that was monitoring nothing. it was just confidently showing me the same snapshot on repeat. what is the dumbest caching bug you have hit in ****?

• 
  Savi (@SaviPabla) reported 21 minutes ago

  @trq212 Please @SlackHQ help us. I’m so tired of telling ppl to download. I’ve now started making videos of the html files and then sharing file as well lol

• 
  jimmy hackett (@j1mmyhackett) reported 21 minutes ago

  the most dangerous bug isn't the one that crashes. it's the one that says 'all good' when it isn't. i had a checker that was supposed to flag a problem. it was reporting green. every single run. turns out it was checking the documentation of the thing it was hunting, not the thing itself. so whenever the problem existed and was documented, the checker saw the doc, said 'found it, it's tracked', and moved on. confident. wrong. this happened twice in the same week in two different places. a report card that shipped one of its four answers unconditionally regardless of the actual data. a monitor that was asserting schema state it had never actually read. both passed. both lied. the lesson i keep relearning: a check that can't fail toward a confident opposite is not a check. it's a placebo. the question isn't 'does this run without error.' the question is 'what would this have to see to fire, and can it actually see that thing.' if the answer to 'can it see it' is 'well, kind of, it reads the notes about it'. you don't have a checker. you have a script that reads its own homework. @SlackHQ kept posting green cards. @AnthropicAI's Claude Code kept summarising them as healthy. everything looked fine. nothing was. what's the quietest false-green you've shipped?

• 
  Theodor Marcu (@theodormarcu) reported 23 minutes ago

  Putting an agent in Slack creates a very different UX problem: it has to know when to shut up We spent a lot of time tuning when Devin should speak, stay quiet, or move work into a separate channel. Devin was the first agent to support Slack natively, so it was great to bring 2 years of those learnings to @SlackHQ for the Slack Code launch.

• 
  Jimidate (@Jimidvte) reported 25 minutes ago

  @SlackHQ, I’m having issues with my mobile notifications. I’ve checked every setting on my phone, uninstalled and reinstalled the app, yet I still don’t receive any push notifications. I’m missing time-sensitive information because I have to open the app before I see any updates.

• 
  Luke Hutchison (@LukeHutchison) reported 27 minutes ago

  Why we're SaaSmaxxing at Perfect Venue Three months ago I tried vibe coding our way out of half our SaaS stack and racked up a $20k Replit bill. Now we are doing the opposite, doubling down on our existing vendors and adding more to our stack. It started small. A better prospecting tool built in Replit. ICP scoring, competitor signals, enrichment. It "worked," and that was the problem. I could come up with an idea and within minutes a version that "worked" was in front of me. It was a lot of fun at first. Once you see how fast you can build exactly what you want, scope creeps hard. Within weeks I was eyeing our ATS, our feedback board, and parts of HubSpot. Part of me also wanted to try vibe coding because when you're leading a company where "apocalypse" is being added to your industry category, I thought it would be good to know if it was a real threat or just hype. We sell software to restaurants and venues. Homegrown spreadsheets and internal tools have always been our biggest competitor, long before AI. So the real question wasn't "can you build your own?" It was: has that math actually shifted, or does it just feel like it has? But then reality set in. I was quickly reminded of the saying in engineering that "the first 90% of a project takes 90% of the time, and the remaining 10% takes another 90% of the time." Getting a demo working was fast, but feature requests and bugs quickly piled up. Obviously we could put the feature requests in the icebox, but that defeats one of the main goals of custom software: being able to customize it to your business. The problem is that customization requires a customizer. Every field and edge case routed through one person. Me. Next, we explored having an engineer work on building a more sustainable solution. However, the direct cost of that engineer working on an internal tool was much higher than all of our SaaS spend combined. The indirect cost was even higher, because it meant that engineer was not working on our core product and building features that could drive much more revenue than the costs they would be able to reduce, not to mention stronger market positioning. There were also other indirect costs. I was interviewing for a recruiting role and explaining to her that we were moving off Ashby to a vibe coded tool. As soon as the words came out of my mouth, I could tell that decision was a bug and not a feature. The reality is that people have worked with many popular SaaS tools for years. They may not love them, but they know how to work with them, and they can focus on doing their job instead of being a design partner for a vibe coded tool. Another challenge was that growth meant building. Even if we were able to build and maintain a custom solution, it was just for a point in time. As the team and company grew, we needed new functionality. With a SaaS vendor, it is very rare that we would need something they didn't already have. They've been solving for the needs of thousands of businesses for years, it's rare we would really need something that another one of their customers had not come across before. Often we'd have to pay more, but it was as simple as a quick upgrade. With a vibe coded solution, it meant understanding the problem, designing and building a solution, and then maintaining it. So where are we today? We're SaaSmaxxing. We're getting everything we already paid for out of the tools we already have - here is how we're doing it. Use AI to better use our SaaS tools. Instead of seeing the problem as AI vs. SaaS, we're using AI to help us better use the tools we have. One of the most frustrating things about SaaS products is that they can be complex and hard to use, with settings buried and functionality hidden. Help centers only go so far. With Claude, I'm able to explain what I'm trying to do, and it can help me configure our existing tools or find a product that integrates with our existing software and solves the problem. Use vibe coding to prototype functionality we wish we had. When there is functionality that we wish we had with our vendors, we'll vibe code a prototype and then find ways to recreate that functionality with our existing tools. Take the 80/20 and move on. Almost every improvement we wanted had a good-enough version in a product we already paid for. Solving for the last 20% was far more expensive than it was worth. Clearly define the problem. As a business owner, my problem was that it was frustrating to log into all of these different tools to understand what was going on with my business. I now use a tool called Viktor to easily pull information across all of our different tools and send reports to me in Slack. File the feature request. Their roadmap is free engineering. A vendor building it once for a thousand customers beats me building it once for us. Has the build-versus-buy line moved? Yes. Building the feature list is dramatically cheaper than it was two years ago. But the feature list was never the expensive part. Owning it is. For the vast majority of our vendors, the amount we pay is a fraction of the value we get, even in the AI era. Our industry is changing faster than ever, and who knows what functionality new models will bring, but this is where we've landed for now. We're buying @HubSpot, @SlackHQ, Canny, Ashby, etc - and focusing on building Perfect Venue faster than ever with the help of AI. Great work @dharmesh and team on the continued updates!

• 
  Calvin Lee (@calvinclee) reported 29 minutes ago

  @SlackHQ the mobile app has significantly regressed. why do I have a DM tab and an activity tab that only lets me filter down to DMs again? I want to easily filter for mentions.