Skip to main content
Case Study · Pods Insurance

A Canadian direct-to-consumer insurance provider

A recurring Apache thread-exhaustion outage had gone unexplained for 6 to 10 months on a LAMP stack handling 2,000+ daily quotes, so this Canadian insurer engaged a 60-hour Flexible Engineering POD. Armakuni preserved diagnostic evidence from the August 1 outage and scoped an observability stack for the 800-queries-per-request pattern, protecting atomic policy binding and payment flows throughout.

1
SOW
60 hours per…
Engagement
$2,400m
AWS funded
4
Outcomes shipped

A Canadian direct-to-consumer insurance provider came to Armakuni for devOps Care POD to diagnose recurring Apache thread exhaustion on a LAMP stack, recommend an observability layer on a previously dark production environment, and deliver reliability fixes that protect atomic policy binding and payment transactions.

The challenge

A Canadian direct-to-consumer insurance provider engaged Armakuni to diagnose and resolve a recurring production outage pattern that had resisted resolution for 6 to 10 months. The 60-hour Flexible Engineering POD brought senior DevOps and full-stack engineers against the production environment under a retainer model A Canadian direct-to-consumer insurance provider can extend into further data and application engineering work.

Scale: Approximately 1,000 policies and 2,000+ quotes processed daily across 2 production servers running a LAMP stack on AWS, with approximately 800 MySQL queries per request.

What we built

Armakuni delivered devOps Care POD to diagnose recurring Apache thread exhaustion on a LAMP stack, recommend an observability layer on a previously dark production environment, and deliver reliability fixes that protect atomic policy binding and payment transactions. The engagement ran as single Flexible Engineering POD, retainer-based, consulting scope. 30 hours allocated upfront, 30 hours reserved for the customer at their discretion. AWS funded the work at $2,400m.

The outcomes

4 measurable outcomes shipped across the engagement. The ones that moved the business the most:

1,000
policies processed daily on the stabilised platform
2,000+
quotes processed daily
800
MySQL queries per request diagnosed and addressed

Diagnostic evidence preserved from the August 1 outage instead of erased by restart. The firm finally has a concrete artifact explaining a production pattern that had been opaque for 6 to 10 months, rather than a guess to defend at every recurrence.

Observability stack scoped for the 800-queries-per-request pattern. RDS Performance Insights at the database layer plus application-level request tracing close the visibility gap that made every previous incident a guessing game.

Reliability recommendations engineered for atomic transaction safety. Connection pooling and session-timeout approaches protect policy binding and payment flows throughout the fix window, improving reliability without transactional risk.

60-hour Flexible Engineering POD established as an ongoing retainer. Remaining hours and follow-on months stay available for further DevOps, database, or full-stack work, structured as a capacity relationship the firm can extend.

Built on AWS

The production environment runs on 4 first-party AWS services. Delivered under our AWS Not specified in source documents. Confirm before publishing. competency.

Amazon RDS for MySQLAmazon RDS Performance InsightsAWS EC2Amazon CloudWatch
AWS Premier Tier Partner
Validated by AWS

What's next

Implementation of the recommended observability stack and reliability fixes. The retainer POD remains available for further data engineering, DevOps, and application engineering work at A Canadian direct-to-consumer insurance provider's discretion.

Want similar outcomes for your platform?

Talk to us about an engagement shaped around the same constraints A Canadian direct-to-consumer insurance provider brought to the table. Most start with an AWS-funded discovery and a conversation with an engineer, not a sales exec.