Ensuring platform stability ahead of post-lockdown demand surge

Preparing for a rapid surge in uptake

As the first national lockdown of 2020 came to an end, the hospitality sector readied itself for a raft of new requirements to ensure venues and processes were Covid-secure. An important aspect of this was providing table service, ideally via contactless ordering.

Our client had a golden opportunity to drive rapid uptake of its Azure-hosted platform in pubs, bars and coffeeshops. But it also had to ensure the application wouldn’t buckle under pressure when demand spiked, especially at weekends.

Initially, the application’s hosting footprint was sized according to the requirements of maximum weekend demand. However, this meant it was over-provisioned for the rest of the week, resulting in a disproportionate monthly fee. General platform stability was also a concern since the company had limited operational resources.

To help resolve these issues, DevOpsGroup was appointed as managed services partner to work alongside the technical team.

“We knew the app would experience a hard and fast spike in demand when pubs, bars and coffeeshops reopened after lockdown. It was vital that the platform was ready to handle this, so we put measures in place to maximise reliability, stability and cost-effectiveness.”

Database optimisation keeps orders flowing

Our cloud engineers provided 24×7 reactive support to ensure emerging issues were fully resolved before they escalated. This was coupled with a Kaizen allocation whereby a set amount of time is ringfenced for continuous improvement on a regular basis.

Since the company needed to achieve a level of operational maturity usually associated with larger organisations, we conducted a Well-Architected Review. While many of the recommended actions weren’t immediately relevant, the outputs provided valuable guidance to ensure decisions were shaped by best practice principles.

When performance issues arose, we quickly identified the root cause and implemented measures to prevent them recurring. We also worked proactively to address typical scale-up concerns – such as database issues – that are associated with performance problems. For instance, noting that some database tables were not being updated on a regular basis, our database specialists conducted bespoke pre-emptive work to rectify the problem. At the same time, we took the opportunity to refactor the index job, improving any indexes with a fragmentation level above 30 percent. Once these measures were in place, performance improvements were quickly realised.

Perhaps most importantly, we enabled the company to automate its provisioning, so the hosting footprint scaled up and down as needed in response to demand. This ensured space was only paid for as and when needed.

Stability, reliability and cost-optimisation

We worked closely with the in-house team to ensure the application could handle the post-lockdown surge in demand. Performance issues were quickly identified and resolved, with no loss of service. Pre-emptive work was also undertaken as needed to avoid further problems. Automated right-sizing enabled the company to optimise cloud costs, cutting its hosting spend by around 30 percent.