In software ops, history isn’t just history—it’s your roadmap. Those years of APM (Application Performance Monitoring) data? They're gold. You can spot seasonal traffic spikes, track down old bugs, or figure out how your systems behave under pressure.
But what if it’s time to change tools?
Maybe your APM costs too much. Maybe the interface is clunky. Or maybe you need better integrations. Whatever the reason, switching APM providers can feel like swapping out your car’s engine while flying down the highway.
And here’s the real kicker: ditching your old tool often means saying goodbye to years of data. So, how do you change APM platforms without losing your history?
Let’s break it down with two real-world stories.
Case 1: E-Commerce Team Cuts Costs, Keeps Data
A mid-sized e-commerce company was paying over $150,000 a year for New Relic. That number stung. So, they made the call to move to Datadog, which had similar features—for about a third of the price.
But here’s the catch: they had three years of data locked inside New Relic’s system. And it wasn’t just fluff. This data showed customer behavior, traffic during holiday sales, and how their infrastructure held up under load.
Walking away from that? Not an option.
So their engineers got scrappy. They used New Relic’s API to pull metric snapshots in batches. Then they reshaped that data to match Datadog’s format.
Here’s a glimpse of how it started:
# Export APM metrics from New Relic
curl -X GET 'https://api.newrelic.com/v2/applications/{app_id}/metrics/data.json' \
-H 'X-License-Key: YOUR_NEW_RELIC_API_KEY' \
-o metrics.json
After exporting the data as JSON, they cleaned it up, mapped the fields, and sent it into Datadog through a custom ingestion pipeline. The process took weeks. They had to watch API rate limits and fix mismatches along the way.
But in the end? It worked. They saved money and kept their data.
Case 2: Financial Firm Plays It Safe with Dual Logging
Now let’s talk about a financial services team. Their challenge? Moving from AppDynamics to Elastic APM—while staying compliant with audit rules. For them, keeping historical logs wasn’t a “nice to have.” It was a requirement.
Instead of bulk-exporting data, they built a dual-logging bridge. It pulled real-time metrics from AppDynamics and sent them to both platforms at once. No gaps. No guesswork.
Here’s the core idea:
# Dual-logging bridge between AppDynamics and Elastic
import time
import elasticsearch
from appdynamics import AppDynamics
appdynamics_client = AppDynamics('YOUR_APPDYNAMICS_CONFIG')
elastic_client = elasticsearch.Elasticsearch(['YOUR_ELASTICSEARCH_URL'])
while True:
metrics = appdynamics_client.get_metrics()
elastic_client.index(index="app_metrics", doc_type="_doc", body=metrics)
time.sleep(15)
Was it more complex? Sure. It used more resources, and there was extra overhead. But for a regulated environment, it was worth every line of code. They stayed compliant, kept visibility, and never missed a beat.
What We Learned
Switching APM tools doesn’t mean starting from scratch. But you do need a plan.
Here’s what came up again and again:
- Start early. Exporting old metrics isn’t always easy. Some APIs have rate limits or weird formats.
- Automate smart. Data transformation scripts need to handle mismatches in units, timestamps, and field names.
- Think about dual-logging. For high-stakes environments, running both systems in parallel is often safer.
- Be ready for ingestion hiccups. Pushing old data into a new system might trigger false alerts or cost surprises.
- Check your coverage. Always validate your imports—don’t assume everything made it over cleanly.
One Last Thought
Historical performance data is more than a record—it’s your context. Losing it is like tearing out the first half of a book and trying to make sense of the ending. But with the right prep, scripts, and strategy, you can switch APM platforms and bring your history with you.
You don’t have to fly blind.