As more organisations base operational decisions based on predictive analytics, understanding how to standardize data collection– identifying and eradicating the incorrect, irrelevant or inaccurate metrics from your stack – will be necessary to help drive business productivity and cost savings, and improve overall application performance. One of the easiest ways to achieve this is by standardising data collection across your IT stack. However this does not come without its challenges.
The biggest challenge most businesses face when beginning to standardise operational data is that most metric collection isn’t standardised to begin with. Companies must establish a base standard before they can begin to move forward. Many times one team as a whole does not have access to all the data. It is often shared with stakeholders, meaning it can be a challenge in itself to get everyone together, which makes it difficult to collect and standardise in an agreeable format.
As organisations transition from on-premises to the cloud and then to multiple public clouds, the differences between the metrics and monitoring platforms in each of these environments makes them easily siloed.
To counteract all of these issues, there are a number of steps that organisations can do to help standardize metric collection. Here are three examples.
Establish a baseline
Identifying a standard is a good place to begin. Despite the fact that we strive for data-driven decisions, data collection is never seen as a priority.
The first step is to get the entire team on board. Creating a task force of people is one way to overcome, determine and agree on the current data collection processes and agree on an appropriate standard for the entire company.
You’ll likely find that there different procedures (scripts, APIs, plugins) and plenty of overlap within the organization. Bringing a team together will ensure that all endpoints are correctly identified and no data will be overlooked when standardising. Data scientists can help ensure the standardisation is done correctly and effectively, taking into account all different types of data.
Select a standard
Once you have a task force at hand, create a plan to get the process moving. Is it easier to select a single source of truth for monitoring each endpoint, or is it easier to centralize on monitoring-integration-as-a-service? The truth is home grown and open-source solutions will work fine at a certain scale and in smaller use cases.
It goes without saying that standardizing your data collection process can impact many parts of your organization, so making sure IT you involve all teams early, regularly evaluate your progress and communicate your successes can be a huge help. Auditing your strategy regularly is critical. Every six months, be sure to take a look at SLAs, MTTR rates and other KPIs to ensure your data collection strategy is helping them trend in the right direction.
We’re just starting to realize the full potential of IT operational data. Going through the process of standardizing your data collection process is a vital part of utilizing it and maximizing its business value.