Running on Slack?

Slack app, webhook or SSO misbehaving? We ship it properly.

Slack looks like a chat tool to users and like a full event platform to anyone who writes against the API. A webhook that worked for a year suddenly 429s, a slash command stops responding within three seconds, or Enterprise Grid SSO starts bouncing users for reasons the dashboard will not explain. We work through the API, the app manifest and the audit log until the real cause surfaces.

  • Since 2015Fixing technology
  • FreeEstimates
  • US teamReal engineers

slack-app.log

still failing

What it is

Slack at a glance

Slack is a workplace messaging platform launched in 2013 by Stewart Butterfield and the remains of Tiny Speck. It has become the default chat tool for a large share of tech companies, and since 2021 has been part of Salesforce. Underneath the familiar channel UI is a serious API: slash commands, interactive messages, Block Kit, Workflow Builder, incoming and outgoing webhooks, Events API and Socket Mode.

The product comes in Free, Pro, Business+ and Enterprise Grid tiers, with SSO, SCIM and audit features gated behind the higher plans. Most teams encounter Slack as a user and discover it as a platform the day somebody writes a slash command that needs to answer in under three seconds, or an admin enables Enterprise Grid and the integrations that worked yesterday stop working today.

Type
Workplace messaging and integration platform
First released
2013
License
Proprietary SaaS, per-seat pricing with Free, Pro, Business+ and Enterprise Grid tiers
Maintained by
Slack Technologies, a Salesforce company
Typically used for
Team chat, incident channels, bot integrations, workflow automation, SSO-gated collaboration

The honest version

Slack pros and cons

No technology is all upside. This is where Slack earns its place and where it tends to cause trouble.

Where Slack is strong

  • The API is thorough and well documentedEvery surface a bot can touch is covered, from slash commands to modals to home tabs, and the reference is reliable enough to build against without reverse engineering.
  • Block Kit makes rich messages tractableStructured blocks replace the old attachment hack, so a bot message can include buttons, selects and inputs without manually crafting mrkdwn strings.
  • Workflow Builder covers no-code use casesNon-engineers can build request forms, approvals and routing without writing an app, which keeps the simple cases out of the engineering backlog.
  • Universal reach at tech companiesAlmost every vendor you integrate with already has a Slack app, which makes it a cheap control plane for alerts, deploys and status.
  • Enterprise Grid for larger orgsMultiple workspaces under one org, with SSO, DLP, retention policies and audit logs centralized for compliance teams.
  • Granular bot permissionsScopes are fine grained and consent is per-workspace, so a limited-scope bot can be installed without handing the integration the keys to the whole org.

Where Slack bites

  • Three-second ack is unforgivingSlash commands and interactive actions must acknowledge within three seconds. Any synchronous work against another API that takes longer breaks the command, and the user sees 'operation_timeout'.
  • Rate limits are per-method and surprisingTier 1, 2, 3 and 4 limits apply differently to chat.postMessage, files.upload and conversations.list. A nightly job that worked for months can start 429ing when the workspace crosses a threshold.
  • Enterprise Grid changes the rulesScopes, token types, event payloads and channel discovery all behave differently once a workspace becomes part of Grid. Apps that worked in Pro often need refactoring to survive the migration.
  • Message deletion and retention is a policy minefieldRetention is set per workspace and per channel, and legal holds interact with Enterprise Grid differently again. Deleting things is both too easy and too hard, depending on who is asking.
  • Costs climb with seat countBusiness+ and Enterprise Grid pricing scales linearly with headcount, and guest accounts are not free. For a growing company, the Slack bill is a real line item.
  • Socket Mode is tempting and not always rightIt skips the public URL, which is convenient for prototypes, and does not survive at the scale or reliability of a proper event subscription with a public endpoint.

Sound familiar?

Common Slack problems we fix

The Slack problems that bring people to us, and what you have once each one is fixed.

slash command took too long (3000ms)

Slash commands time out before replying

A command that calls another API synchronously and misses the three-second ack. The fix is to ack with an empty response immediately and finish the work with response_url after the fact.

Commands ack fast, reply correctly

ratelimited (Tier 2)

The bot is hitting rate limits

A broadcast that posts to every channel at once, or a backfill job calling conversations.history without pagination awareness. Shaping traffic to the tier limits fixes it in code rather than by upgrading the plan.

Traffic smoothed, limits respected

missing_scope needed: chat:write.customize

An action returns 'missing_scope'

A new feature added to the app manifest that was never re-authorized in the workspace, or a scope the vendor silently deprecated in a recent SDK bump.

Scopes correct and installed

SAML: InResponseTo mismatch

Users bounce out of SSO login

A clock skew between Slack and the identity provider, a certificate rotation nobody announced, or an IdP that is returning a different NameID than the SCIM directory expects.

SSO stable, logins quiet

workflow step failed: variable 'email' not found

A Workflow Builder flow stopped working

A trigger payload that changed shape, a step referring to a variable that was renamed, or a webhook step whose destination returned a non-2xx and silently failed the entire flow.

Workflows running, errors visible

event verification failed, ngrok offline

Event subscriptions keep failing

A development URL that expired, a Lambda that cold-starts past the 3-second verification window, or a WAF rule stripping the x-slack-signature header.

Events delivered and verified

channel_not_found

The bot cannot post to a channel

A private channel the bot was never invited to, a channel archived under Grid, or an org-shared channel that requires a different scope than the one granted.

Posts land in the right channel

What we do

Slack help, apps and admin

Ring us for one failing webhook, or let us run every Slack integration you have.

Rescue
App returning 429s, slash command stuck on 'operation_timeout', or SSO bouncing users at login. We stabilise the integration first and write up the real cause second.
Slack apps and bots
Slash commands, Block Kit modals, event subscriptions, home tabs and interactive messages built to Slack's actual guidelines, not a demo tutorial.
Webhooks and alerts
Incoming webhooks for alerts, outgoing routing to on-call tools, and the kind of rate-limit handling that survives a real incident rather than quietly throttling.
Enterprise Grid migrations
Moving workspaces into Grid without breaking the integrations, re-scoping the apps, and taking the admin team through what changes about SSO, retention and channel discovery.
Workflow Builder cleanup
Flows audited for who owns them, broken steps repaired, and the critical ones rewritten in code when the UI reaches its limit.
Managed Tech
Ongoing watch on API error rates, webhook health, scope drift and the audit log, with a named engineer who can respond when the integration is the incident.

How it works

From first message to fixed

  1. Look Free estimate

    Tell us what's going wrong. We look at it, then tell you plainly what's wrong and what it will cost.

  2. Fix

    We fix the cause, not the symptom. If something is down, getting it back up comes first.

  3. Harden

    Backups, monitoring, updates and access, so the same problem can't quietly come back.

  4. Hand back or stay on

    You get a plain-language write-up. Or we stay on for hosting, upkeep and new work.

Often found together

Running Slack alongside something else?

We work on the whole stack, so the problem can be wherever it actually is.

Questions

Slack questions

Something else? Call+1 737 727 7327

Why is my Slack slash command returning 'operation_timeout'?

Because the handler did not reply with any response within three seconds. Slack considers that a failure and tells the user the command timed out. The right pattern is to acknowledge immediately with a 200 and an empty body, then do the real work in the background and post the result back to the response_url. Doing all the work inside the first request only works for the fastest APIs.

Why does our bot keep hitting Slack rate limits?

Each method has its own tier, and chat.postMessage, files.upload and conversations.list all throttle differently. The usual pattern that causes trouble is a job that iterates channels or users with no backoff, or a broadcast that posts to every channel at once. We add respect for the Retry-After header and shape the traffic to the tier so the limit stops being hit in the first place.

We are moving to Enterprise Grid, what breaks?

Tokens change shape, some scopes behave differently, org-wide apps replace workspace-scoped ones, and event payloads can include new fields your handlers were not written to expect. We take each installed app, map it to the new world, and migrate them in a planned sequence rather than all at once on the cutover day.

How do we fix SSO users bouncing out of Slack?

Usually a SAML clock skew, a changed certificate at the IdP, or a NameID that does not match what SCIM provisioned. The audit log and the identity provider's log together show which. We walk through both sides, fix the mismatch, and tighten the configuration so a quiet certificate rotation does not take the whole org offline again.

Should we use Socket Mode or a public URL for Slack events?

For a small internal tool, Socket Mode is fine and lets you skip hosting a public endpoint. For anything you expect to be reliable under load or shared across workspaces, a proper event subscription with a signed, public URL is more robust and much easier to monitor. We pick the one that matches the use case rather than the one that is quickest to demo.

How much does Slack help cost?

The quote is free. Share the app manifest or a short description of what the bot is meant to do and where it is failing, and we come back with a scope and a fixed price before touching the workspace.

Free estimate

Slack giving you trouble?

Apps, webhooks, workflows, SSO and Enterprise Grid. If it runs on the Slack platform, we have shipped and fixed something like it.

  • No charge to find out what it will take
  • If it's urgent, say so

Replies come from a person. By sending, you authorize us to email you back. No marketing.