Production monitoring: errors, performance and logs
We set up error tracking, performance monitoring and structured logging on the business systems we look after, so problems are seen and fixed before users have to report them.
What you probably want to know first
- How would we know about problems before users tell us?
- Error tracking records every unhandled error in the browser and on the server with the context needed to fix it. Performance monitoring shows which pages and queries are slow and when that changed. Uptime checks alert when the system stops responding. Together they turn most user complaints into something already being worked on.
- Does monitoring need changes to our application?
- Small ones. Error tracking and performance agents are usually a package and a few lines of configuration. Useful logging takes longer, and is added as areas of the code are worked on.
- What will it cost?
- The tools have free or low-cost tiers that cover many business systems. We size the plan to the volume of errors and data you actually generate, and filter out noise so you do not pay to store it.
- What have you done with systems like this?
- OpenIRIS is monitored with Sentry for errors in the browser and on the server, New Relic for application performance and uptime checks, and Seq for structured logs; new errors are triaged and turned into fixes. Logtek's system is monitored with New Relic.
Who this page is for
This page is for IT managers who find out about problems with a business system from its users, and want to know first.
Why it matters for support
Support without monitoring is reactive: a user notices a problem, works around it for a while, then reports it with whatever detail they remember. By then the error may have happened hundreds of times. With monitoring, the same problem shows up the first time it happens, with the page, the user’s browser, the server’s stack trace and the release that introduced it. That is the difference between fixing something in an hour and spending a day trying to reproduce it.
How we use it
On the systems we support, new errors are triaged as they appear. Each is fixed, scheduled, or explained and silenced if it is harmless, so the list stays short enough to read. Performance data decides where tuning effort goes: we measure where the time is spent before changing any code.
Talk to us
Describe the system and what you need. You will hear back from someone who can answer technical questions.
Discuss monitoring for your system 0800 433 7990What we set up
- Error tracking
- Every unhandled error in the browser and on the server, grouped, counted and linked to the release that introduced it.
- Performance monitoring
- Response times for each page and API call, the database queries behind them, and alerts when they slow down.
- Uptime checks
- Automated checks from outside your network that sign in and use the system, and raise an alert if they cannot.
- Structured logging
- Searchable logs that tie a user's request to everything the server did with it.
- Regular review
- New errors are looked at as part of support and either fixed, scheduled or explained.
- Alerts that matter
- Alerts tuned so that each one needs action, sent to the people who can act.
How we add monitoring to an existing system
Error tracking first
Usually the quickest win: within days you can see which errors users hit and how often.
Performance next
An agent on the servers shows the slowest pages and queries, measured rather than guessed.
Uptime checks
Scripted checks of sign-in and the main workflows, run every few minutes.
Logging as we go
Structured logs added to each area as it is changed.
Tune
Noise filtered out and alerts adjusted until each one is worth reading.
A good fit when
- Users report errors before anyone in IT knows about them.
- The system is slow at times and nobody can say why.
- Logs exist but nobody can search them or connect them to a user's problem.
Probably not for you if
- You need infrastructure monitoring for an office network or desktops; we monitor business applications and the servers they run on.
Questions we are asked
Does error tracking send our data to a third party?
It sends error details to the monitoring service. We configure it to strip personal data before sending and choose a service region that fits your data protection requirements.
Can you use the monitoring tools we already have?
Yes. If your organisation already uses a monitoring product, we work with it rather than adding another.
Tell us about your system
Say what it does, what it is built on and what is worrying you. We will reply with what we would look at first and whether we are the right people to help.