Find the outcome you need.
Start with a symptom, a platform or the thing you want working.
With JavaScript, search filters the whole collection in your browser. Without it, use the linked pages and section navigation below; query filters need JavaScript.
Showing entries 641–680 of 775 in this first collection.
Platform
SQLite in a growing app: loose types, optional foreign keys and one writer at a time
What SQLite does differently from PostgreSQL or MySQL, how to check a SQLite database for integrity problems, and when an app has outgrown it.
Collection
A slow or failing database-backed request: find the first failing layer, in this order
An ordered process map from "the page is slow" to one named cause: query count, one statement, indexes, waits, connections, cache and data shape, each linked to its guide and fixed-scope job.
Collection
A background job or schedule misbehaves: did it run, did it run twice, did it finish?
An ordered process map for scheduled jobs and queue workers: whether the run happened, why it did not, whether it repeated, where failed messages went, and who is told, each linked to a guide and a fixed-scope job.
Troubleshooting guide
Bookings that duplicate or land an hour out in Google Calendar: event identity and time zones
Give each booking a stable Google Calendar event ID so retries never duplicate it, and store the time zone so events show the right local hour across clock changes.
Troubleshooting guide
Why a Google Calendar connection stops working after a week, or months later
The documented reasons a Google refresh token stops working, and how to design a visible reconnect state so bookings are never lost while it is broken.
Troubleshooting guide
Address autocomplete on a checkout: which Google component, which key settings, which fields
Use the current Places autocomplete component, restrict the key to your site and to the Maps JavaScript API and Places API (New), and request only the address fields the form uses.
Troubleshooting guide
Adding Sign in with Google to existing accounts: key on the subject ID and choose a linking rule
Why the provider's subject identifier, not email, identifies a Google user, and three linking rules with the risk each carries.
Troubleshooting guide
Google sign-in callback checks: state, nonce, PKCE, one-time code and ID token validation
The checks a sign-in callback must make before it starts a session, including PKCE, and a test for each failure using synthetic tokens and a stand-in token endpoint.
Inspectable example
Synthetic matrix: Twilio status callbacks arriving in the wrong order, and the stored result each must give
Invented message callback sequences show a forward-only status rule keeping a final state when older statuses arrive late, with the cases an acceptance test should include.
Inspectable example
Synthetic test vectors for a Slack events endpoint: signature, five-minute window and repeated events
Five invented requests with their computed signatures show which a verifier should accept and which it should reject, and why an event ID must be claimed before the work is queued so a repeat or a simultaneous duplicate does not run it twice.
Inspectable example
Synthetic retry traces: waits, attempt limits, a Retry-After date and a timed-out create
Worked arithmetic for a capped, jittered retry policy with invented random draws, plus the records a safe retried write should leave.
Inspectable example
Synthetic booking table: Google Calendar event IDs and local times across clock changes
Five invented bookings show the event ID, zone and UTC offset a sync should write, including a local time that does not exist and one that happens twice.
Buyer collection
For a founder connecting one outside service to an app: ask for one result you can test
Turn 'connect it to Twilio, Slack or Google' into one named result with a test, decide who holds the credentials and pick the single outcome that fits.
Buyer collection
For a small dev team with several outside services: one retry fix, one module or a monthly check
Choose between hardening one failing client, consolidating several services behind one module and keeping a register of credentials, quotas and deprecations.
Collection
A notification did not arrive: check these in order before changing any code
An ordered map from 'the customer says it never came' through the app's own record, the provider's delivery trail, the callback, the recipient's state and the sender set-up, with the guide for each step.
Collection
Accepting a new integration: the order in which to test it before you trust it
An ordered acceptance map for any new connection to an outside service: result, credentials, test route, happy path, failure paths, inbound verification, retries and monitoring.
Troubleshooting guide
SendGrid says the domain is verified, but is your mail actually aligned with your From address?
Read a received message's headers to tell whether SendGrid's SPF and DKIM results match your From domain, and what to check when they do not.
Troubleshooting guide
Bounced, blocked or dropped: what SendGrid's event feed tells your app to do
Separate permanent bounces, temporary blocks and drops caused by suppression, and store each once so your app stops mailing addresses that cannot receive.
Troubleshooting guide
Verifying SendGrid's signed event webhook: why the raw body matters
How SendGrid signs its event feed, why re-serialised JSON breaks verification, and how to test that a forged event flags nobody.
Troubleshooting guide
Twilio says sent, then delivered, then sent again: store a final state, not the latest arrival
Rank Twilio message statuses so an out-of-order callback never overwrites a final result, and keep error codes as information rather than rules.
Troubleshooting guide
Twilio callback checks fail or never fire: the exact URL and what test credentials cannot do
Why Twilio's request signature depends on the exact URL it called, and why test credentials cannot show delivery callbacks at all.
Troubleshooting guide
Alerting Slack from your app: webhook or bot token, and how to avoid slowing or flooding it
Choose between an incoming webhook and a bot token, keep alerts outside the customer's request, and respect Slack's per-channel rate limit and Retry-After.
Troubleshooting guide
Slack events hit your endpoint twice, or not at all: signature, timestamp and the three-second rule
How Slack signs events, why the raw body and a five-minute window matter, and how to reply in time without running the work twice.
Platform
SendGrid in your app: authenticated sending, then what the delivery events tell you
Understand what SendGrid's domain authentication proves, what its event feed reports and which parts of a transactional email connection are yours to build and test.
Platform
Twilio SMS in your app: sending is one call, knowing what happened is the work
Understand message statuses, out-of-order delivery callbacks, request signatures and sender registration before connecting an app to Twilio messaging.
Platform
Slack API in your app: posting alerts and receiving events are two different jobs
Understand when to use an incoming webhook or a bot token, how Slack rate limits posts, and the signing, timing and retry rules for events it sends to your app.
Troubleshooting guide
Direct uploads with signed storage links: what the link allows, how long it lasts and what it does not check
A signed upload link is a bearer token tied to the credentials that made it; design the key, lifetime, after-upload checks and the cleanup of uploads nobody confirms around that, and note where S3-compatible stores differ.
Troubleshooting guide
Rate limited or timing out: wait as the provider asks, then back off with spread
Read Retry-After in both of its forms, add capped exponential backoff with randomness, bound the attempts and total time, and test the policy with a scripted fake provider.
Troubleshooting guide
A timed-out create may have worked: making retried writes safe
Three ways to avoid duplicate records when retrying a create, from provider idempotency keys to lookup and an attempt ledger, and what to do when none is possible.
Troubleshooting guide
Website form to Pipedrive: store first, match by email, create a linked lead, respect the limits
The order of calls, the API versions, the matching rule, the rate limits and the outage design for pushing a form enquiry into Pipedrive without losing it, and the lookup that keeps a lost reply from creating a second lead.
Troubleshooting guide
Keep a register of your integrations: what expires, what limits you and what gets retired
A one-page register of each outside service's credentials, owners, dates and limits, with documented examples of what ends and where to watch for notices.
Troubleshooting guide
Certbot renewal on your own server: dry run, schedule, reload hook and the old 30-day rule
How to check that Let's Encrypt renewal will work before the certificate expires, what actually schedules it, why the web server must reload, and why the renewal threshold changed.
Troubleshooting guide
Certificate validation fails: port 80, redirects, DNS challenges and rate limits
What the HTTP-01 challenge needs from your server, when the DNS-01 challenge is the right choice, and how failed attempts can lock you out for an hour or a week.
Troubleshooting guide
Terraform: read the plan like a diff, keep secrets out of state, and protect what holds data
How to read a plan before applying, what a saved plan file and state can expose, when a module is worth writing, and the limits of prevent_destroy and import.
Platform
Terraform: write the resources as code, review the plan, and keep state safe
What a Terraform module job covers, what a monthly drift review does, and the acceptance evidence for each: a reviewed plan, a clean post-apply plan and a state you can protect.
Inspectable example
Worked example: synthetic Terraform plan review notes
A labelled synthetic example: two plans of one small module in the real Terraform output format, with the reviewer's notes beside the lines that matter, showing how creates, a replacement, an in-place change and a sensitive value are explained before anyone applies.
Troubleshooting guide
Kubernetes pod will not start: CrashLoopBackOff and ImagePullBackOff are different problems
How to read container state, previous logs and events to tell a crashing container from an image that cannot be pulled, and which manifest or values change fits each.
Troubleshooting guide
Pods Running but not Ready, and rollouts that never finish: probes and the progress deadline
What liveness, readiness and startup probes each do, why a liveness probe can cause an outage, how a Deployment reports a stalled rollout, and how to roll back safely.
Platform
Kubernetes: find out whether the manifest, the image or the cluster is at fault
How to separate a Deployment problem you can fix in a manifest or Helm values from application bugs and cluster faults, and how a one-Deployment repair is accepted.
Troubleshooting guide
Moving a PostgreSQL database to a new server: dump, roles, restore and the cutover freeze
What pg_dump does and does not include, how to rehearse a restore, how roles and ownership affect it, and why writes must stop before the final dump.