Operations | Monitoring | ITSM | DevOps | Cloud

In 2023, the global regulatory fines exceeded a colossal $10.5bn. It is not an isolated story. For the past few years, data, privacy, and industry-specific regulations have been getting stricter, enforcement is becoming rigorous, and non-compliance fines are going through the roof. Just look at this list on CSO Online of the biggest data breaches and subsequent fines companies like Meta, Amazon, and Equifax experienced in recent history.

Telemetry data sent from applications often contains Personally Identifying Information (PII) like names, user IDs, phone numbers, and other information that must be obfuscated before the data is sent to storage or observability tools, in order to be in compliance with corporate or government policies such as HIPAA in the US or the GDPR in the EU.

At the most abstract level, a data pipeline is a series of steps for processing data, where the type of data being processed determines the types and order of the steps. In other words, a data pipeline is an algorithm, and standard data types can be processed in a standard way, just as solving an algebra problem follows a standard order of operations.

OpenTelemetry (OTel) is an open-source framework designed to standardize and automate telemetry data collection, enabling you to collect, process, and distribute telemetry data from your system across vendors. Telemetry data is traditionally in disparate formats, and OTel serves as a universal standard to support data management and portability.

Here at Mezmo, we see the purpose of a telemetry pipeline is to help ingest, profile, transform, and route data to control costs and drive actionability. There are many ways to do that as we’ve previously discussed in our blogs, but today I’m going to talk about real-time alerting on data in motion, yes - on streaming data, before it reaches its destination.

This is going to be the first of at least two posts about how we, in the Mezmo platform team, use Pipeline to handle metrics. It’s an ongoing effort and everything might change as we move forward and learn, but this is our plan and vision.

In our webinar, Mastering Telemetry Pipelines: A DevOps Lifecycle Approach to Data Management, hosted by Mezmo’s Bill Balnave, VP of Technical Services, and Bill Meyer, Principal Solutions Engineer, we showcased a unique data-engineering approach to telemetry data management that comprises three phases: Understand, Optimize, and Respond.

Imagine a well-designed plumbing system with pipes carrying water from a well, a reservoir, and an underground storage tank to various rooms in your house. It will have valves, pumps, and filters to ensure the water is of good quality and is supplied with adequate pressure. It will also have pressure gauges installed at some key points to monitor whether the system is functioning efficiently. From time to time, you will check pressure, water purity, and if there are any issues across the system.

I recently attended SRECon in San Francisco on March 18 - 20, a show dedicated to a gathering of engineers who care deeply about site reliability, systems engineering, and working with complex distributed systems at scale. While there were a lot of talks, I’ll focus on a few areas that gave me the most insight into how having the right data impacts an SREs and an organization’s success.