Data Platform Support

A data platform does not run itself. Pipelines fail at night, refreshes stall, capacity costs drift upwards, and upgrades wait because the team who built it has moved on to new projects. Often it falls to one or two people who cannot take a holiday. A managed service puts a team behind the platform with set response times, clear reporting, and someone who is responsible for keeping it healthy.

Problems are found before users find them

Monitoring checks pipelines, data freshness, and platform capacity, and alerts our on-call engineer first.

Response times in writing

Each severity level has an agreed time to respond and to resolve, with reporting against it every month.

The platform keeps improving

A share of every month goes to fixing root causes, trimming cost, and tidying what has built up.

Monitoring and alerting

Pipeline status, data freshness, report refreshes, and platform capacity watched continuously, with alerts that mean something.

Incident response

Triage, fix, and written follow-up for every incident, with the cause found and recorded.

Patching and upgrades

Regular, scheduled updates to platforms, databases, and tooling, tested before they reach production.

Cost and performance tuning

Monthly review of spend and slow queries, with changes made and the saving reported.

Backup and recovery

Backups checked regularly and a recovery test run each quarter, so you know they work.

Change requests and small projects

A monthly allowance for new sources, new reports, and configuration changes, handled through one queue.

• Critical: the platform is down or data is wrong in a way that affects the business

  What we do

  An engineer picks it up straight away, keeps you updated, and works until it is fixed or a workaround is in place.

  What you can expect

  Response within 30 minutes, any time of day on the 24/7 tier, with updates every hour.

• High: an important pipeline or report is failing

  What we do

  We find the cause, fix the data, and rerun what is needed so reports are right by the next business day.

  What you can expect

  Response within 2 hours, with a fix or a workaround the same day.

• Normal: something is slow, odd, or needs a small change

  What we do

  It goes into the shared queue, is ranked with you, and is worked in order.

  What you can expect

  Response within 1 business day, handled within the monthly allowance.

• Planned: patching, upgrades, and tuning

  What we do

  We schedule the work with you in advance, test it first, and run it outside peak hours.

  What you can expect

  No surprises, and a short summary after each change.

• Review: how is the platform doing overall?

  What we do

  A monthly meeting covering incidents, cost, performance, and what we plan to improve next.

  What you can expect

  A written report, and a quarterly look at whether the service level still fits.

1. Phase 1

   2 weeks

   Discover

   • Review your architecture, pipelines, and workspaces with the people who built them
   • List the known problems and the checks your team runs by hand
   • Agree the service tier, severity levels, and escalation contacts

   You receive

   Service scope, contact list, and a list of risks found.

2. Phase 2

   2 weeks

   Set up

   • Install or connect monitoring and alerting
   • Write runbooks for the common failures
   • Set up access for our engineers with the least privilege needed

   You receive

   Live monitoring, runbooks, and an access review.

3. Phase 3

   4 weeks

   Run alongside

   • Handle alerts together with your team
   • Fix the repeat problems that appear
   • Check that response times match the agreement

   You receive

   Tuned alerts and a go-live decision signed off by you.

4. Phase 4

   Ongoing

   Operate and improve

   • Take full responsibility under the agreed service levels
   • Report each month and meet to review it
   • Use a fixed share of time to fix causes, not just symptoms

   You receive

   Monthly service report and a quarterly improvement plan.

How we look after your environment

Least-privilege access

Our engineers get only the access the work needs, through named accounts that are reviewed every quarter.

Every action logged

Changes to your environment are recorded with who made them and why, and you can read the log at any time.

Change approval

Planned changes follow the approval route you set, and urgent fixes are reported to you straight afterwards.

Written runbooks

Everything we do is documented in your own repository, so you are never dependent on our memory.

Honest reporting

The monthly report shows every incident and every missed target, along with what we did about it.

Tools we use

Monitoring

Fabric capacity metrics, Databricks system tables, Azure Monitor, Datadog, Grafana

Incident management

PagerDuty, Opsgenie, ServiceNow, Jira Service Management

Data pipelines

Fabric Data Factory, Databricks Workflows, Azure Data Factory, Airflow, Dagster, dbt

Platforms

Microsoft Fabric, Databricks, Snowflake, Azure Synapse, Power BI

Release and configuration

Git, Fabric deployment pipelines, Terraform for Databricks, Azure DevOps, GitHub Actions

Reporting

Monthly service report, dashboards for uptime, cost, and response times

We use your monitoring tools where you have them, and set them up where you do not.