voyager624 - Fotolia
A week after going public, Slack's collaboration service was partially down for more than 12 hours, leaving some customers unable to send messages or perform other basic tasks.
The Slack outage started on Friday when some of its servers "became unavailable," the company said. Matters grew worse later in the day when "a separate issue increased server load."
On Monday, Slack declined to explain why those servers became unavailable or to clarify what the separate issue was. The company relies heavily on the cloud infrastructure of Amazon Web Services, but it would not say whether the blame rested with the AWS or Slack platform.
At its peak, the issue thwarted 10% to 25% of attempts to send messages, upload files and perform basic tasks. And for a half-hour period in the afternoon, some users were unable to access Slack at all. In all, service interruptions persisted from 4:30 a.m. PDT to 7:20 p.m. PDT on Friday.
"Microsoft and Cisco have both had outages in recent months, as well, so I don't think there's a particular issue with Slack," said Irwin Lazar, analyst at Nemertes Research in Mokena, Ill. "But it does reinforce that organizations need backup plans to enable them to continue to communicate and collaborate in the event of cloud provider outages."
Last week's daylong glitch came one week after Slack went public in one of the tech industry's most anticipated public offerings of the year.
In a regulatory filing ahead of its public listing, Slack warned potential investors of financial risks related to service incidents like the one it endured last week. Outages damage Slack's reputation and can cost the company customers, the filing said.
"Our continued growth depends, in part, on the ability of existing and potential organizations on Slack to access Slack 24 hours a day, seven days a week, without interruption or degradation of performance," Slack wrote.
Slack's stock was down 1% in extended trading on Friday following the outage, Reuters reported. Despite the incident, Slack has maintained uptimes above 99.9% every month for the past year.
Issues more common than Slack outage calendar suggests
Last summer, following a string of crashes that drew media attention, Slack redesigned its status website to make it harder to get a bird's-eye view of its connectivity issues. The calendar used to display a yellow or red indicator on every day that users reported service issues, regardless of severity.
Now, the calendar only flags the most significant incidents with a red outage symbol on the calendar. Slack still reports other events, but users -- or prospective customers -- must click through each day on the calendar to get a count of them.
For example, at first glance, the 2019 calendar makes it seem like there have been only five incidents this year, because only five days are marked with red outage symbols. However, clicking through each day on the calendar reveals that users have reported more than 60 other, less significant service issues this year.
Friday's incident will raise common questions about the reliability of cloud services, potentially giving a boost to on-premises products and other Slack competitors, said Art Schoeller, analyst at Forester Research. But in-house systems suffer outages, as well, he said.
"Slack, and team messaging, continues on a roll," Schoeller said. "The on-prem crowd will increasingly become a smaller and smaller 'corner case,' despite these outages. And there will be more."