1. Home
  2. Guides
  3. Implementation
  4. Migrating .NET Apps from Stackify Retrace to Atatus
IntermediateImplementation

Migrating .NET Apps from Stackify Retrace to Atatus

Replace the Stackify Retrace .NET profiler with the Atatus .NET agent on IIS, .NET Framework and .NET 5+, one app at a time, so every app stays monitored until Retrace shuts down on March 31, 2027.

10 min readAtatus TeamUpdated November 16, 202610 sections

01Before you start: list your Retrace .NET apps

Every Retrace profiler has to be replaced. A short inventory tells you how.

Stackify Retrace shuts down on March 31, 2027, and BMC says there is no automated way to move Retrace data or setup to another tool. Each Retrace .NET profiler has to be replaced by the new tool’s agent. This guide does that one app at a time, so every app is always watched by one of the two tools.

Runtime: .NET Framework 4.x, .NET Core 3.1 or .NET 5 to 9. This decides which Atatus install method you can use.

Hosting: IIS (note the application pool), a Windows service, Linux, Docker or Kubernetes.

Bitness: the Atatus profiler supports 64-bit processes only. On IIS, check that 'Enable 32-Bit Applications' is off for the pools you want to instrument with it.

What your team uses in Retrace: traces, errors, logs, server metrics, alerts and dashboards. These are what you rebuild in Atatus.

Export the logs and dashboards you must keep before you change anything. See how to export Stackify Retrace data. For a feature-by-feature comparison, see Atatus as a Stackify alternative.

02What Atatus gives you for .NET

Where the things you use in Retrace live in Atatus.

Request traces: incoming requests in ASP.NET, ASP.NET Core and Kestrel, and outgoing HTTP calls, are traced automatically. See .NET APM.

Database calls: SQL Server (SqlClient), MongoDB and Redis queries show up inside each trace. For the SQL Server itself, see SQL Server monitoring.

Messaging: Azure Service Bus calls are traced too.

Logs: ILogger logs go straight from the agent to Atatus, linked to the request that wrote them. Serilog and NLog have their own Atatus sink and target. See .NET log capture.

Servers, alerts and dashboards: infrastructure monitoring, alerting and dashboards.

OpenTelemetry: if you would rather use the OpenTelemetry .NET SDK than an agent, Atatus accepts OTLP. See moving from Retrace to OpenTelemetry.

03Pick an install method: profiler or NuGet

Both work. The profiler needs no code changes, like Retrace.

Profiler (no code changes): you set environment variables and the profiler attaches when the app starts. It works for .NET Framework 4.x and .NET 5 to 9, on Windows and Linux, 64-bit only. This is the closest match to how Retrace works.

NuGet package: you add Atatus.AspNetFullFramework (.NET Framework) or Atatus.NetCoreAll (.NET Core and .NET 5+) plus a few lines of config, then redeploy. Apps on .NET Core 3.1 must use this method, because the profiler does not support them.

One profiler per process: a .NET process can only have one CLR profiler attached. The Retrace .NET profiler is one, and the Atatus profiler is another, so the two cannot run in the same app at the same time. The Atatus profiler docs say to disable any other APM profiler before attaching it.

04Step 1: Pilot one app

Prove the setup on something low-risk before you touch the rest.

Pick an app with steady traffic and low risk, such as an internal tool or one server of a load-balanced site. Because two profilers cannot share a process, run the pilot where Retrace is not attached: a staging copy of the app, or one server where you remove the Retrace agent while the other servers keep reporting to Retrace.

Compare the two tools for a few days. Request rate, response times, error count and the slowest endpoints should line up. Small differences are normal, because each tool groups and samples requests its own way.

05Step 2: IIS and .NET Framework apps

On IIS the Atatus profiler is set per application pool.

Download the Windows profiler, extract it to a folder such as C:\atatus, and give the application pool identity read and execute access to that folder.

Windows blocks DLLs downloaded from the internet. Right-click atatus_profiler.dll, open Properties and select Unblock. If you skip this, the profiler can fail to attach without showing any error.

Set the environment variables on each pool with the PowerShell script in the Atatus IIS install guide. .NET Framework pools use the COR_ variables and .NET Core pools use the CORECLR_ variables. Never mix the two.

IIS 10 or later is recommended, because it can scope environment variables to one pool. On IIS 7 to 8.5 they have to be set for the whole server, which changes every pool at once. Run the pool in Integrated pipeline mode.

Recycle the pool to start the agent, or run the recycle script later in a maintenance window.

Prefer NuGet for a .NET Framework app? Install Atatus.AspNetFullFramework, add your app name and license key to Web.config, register AtatusModule and redeploy. The .NET Framework install guide has the exact config.

06Step 3: .NET 5+, .NET Core, Linux and Docker

Same agent, with the CORECLR_ variables or the Atatus.NetCoreAll package.

Profiler: on Linux or Windows, extract the profiler and set the CORECLR_ variables before the app starts. Supported on .NET 5, 6, 7, 8 and 9.

NuGet: add Atatus.NetCoreAll, then call builder.Services.AddAllAtatus() in Program.cs on .NET 7 and above, or app.UseAllAtatus(Configuration) in Startup.cs on .NET Core 3.1 to .NET 6. Put the app name and license key in appsettings.json.

ASP.NET Core on IIS: it runs behind the ASP.NET Core Module, so the same NuGet setup works. For the profiler on IIS, set the CORECLR_ variables on the pool as in step 2.

Docker and Kubernetes: the profiler can be built into your image. The Atatus profiler docs link to the Docker setup, and the .NET Core install guide covers the rest.

07Step 4: Turn on log capture

Get your logs next to the request that wrote them.

ILogger: set CaptureLogs to true in the Atatus section of appsettings.json, or set the ATATUS_CAPTURE_LOGS=true environment variable. Profiler installs are configured through environment variables, so use the variable there. Use agent 3.2.2 or later, which also captures logs in profiler mode, including .NET Framework 4.6.2+ apps on IIS.

Log lines written during a traced request carry that trace and transaction, so you can jump from a log line to the slow request and back. Error and Critical log events also appear in error tracking.

Serilog and NLog: add Atatus.Serilog.Sinks with .WriteTo.Atatus(), or Atatus.NLog.Targets with an Atatus target. Both reuse the agent’s license key and app name. Details are in the .NET log correlation docs.

08Step 5: Rebuild alerts and dashboards

Start with the alerts that wake someone up.

Use the list of alert rules you wrote down from Retrace. Rebuild the ones that page people first: high error rate, slow responses on key endpoints, and servers running low on memory or disk. Set them up in Atatus alerting.

Wait until Atatus has a few days of data before you set thresholds, so they match what normal looks like in the new tool.

Recreate the views each team checks during an incident in Atatus dashboards, most important first.

09Step 6: Move the rest in waves, then remove Retrace

Small groups, each one checked before the next.

Move apps in small groups. For each app, remove the Retrace profiler, attach Atatus, check that traces and logs arrive, and confirm its alerts reach the right people.

Keep Retrace on an app until Atatus covers it, and aim to finish well before March 31, 2027 so there is time to fix surprises.

When the last app has moved, uninstall the Retrace agent from each server and check that no COR_PROFILER or CORECLR_PROFILER setting still points at the Retrace profiler.

10Troubleshooting: no data after install

The usual causes, quickest checks first.

Another profiler is attached: only one CLR profiler can attach to a process. Look for a leftover COR_PROFILER or CORECLR_PROFILER value from Retrace on the pool, service or server.

The DLL is blocked: unblock atatus_profiler.dll in its file properties and restart the app.

Wrong variable prefix: COR_ is for .NET Framework, CORECLR_ is for .NET Core and .NET 5+.

Folder permissions: the IIS pool identity needs read and execute access to the profiler folder.

A 32-bit process: the profiler supports 64-bit only.

A .NET Core 3.1 app: use the NuGet package, because the profiler does not support it.

Still nothing: set ATATUS_LOG_LEVEL=Debug, restart, and check the agent log as described in the Atatus .NET troubleshooting docs.

FAQ

Frequently asked questions

Talk to an engineer
  • Yes. Atatus supports .NET Framework 4.x apps on IIS, either with the profiler (no code changes, set per application pool) or with the Atatus.AspNetFullFramework NuGet package.

  • Not with both profilers. A .NET process can only have one CLR profiler attached. Pilot Atatus on a staging copy or on a server without the Retrace agent, then remove the Retrace profiler from each app as you move it.

  • Yes. With .NET agent 3.2.2 or later, turn on CaptureLogs and Atatus captures ILogger logs linked to the request that wrote them, including in profiler mode for .NET Framework 4.6.2+ apps on IIS. Serilog and NLog have their own Atatus sink and target.

  • The profiler supports .NET Framework 4.x and .NET 5, 6, 7, 8 and 9 on 64-bit Windows and Linux. The NuGet package also supports .NET Core 3.1.

  • Yes. Atatus accepts OpenTelemetry data over OTLP, and the Atatus docs show how to set up the OpenTelemetry .NET instrumentation to send traces to Atatus.

Get started today

Monitor your applications with Atatus

Put the concepts from this guide into practice. Set up full-stack observability in minutes with no credit card required.

  • No credit card required
  • 14-day free trial
  • Setup in minutes
Keep reading
ImplementationBeginner

How to Export Your Stackify Retrace Data Before March 31, 2027

Stackify Retrace shuts down on March 31, 2027, and its data does not move to any other tool. Here is what you can export, the limits, what you have to write down by hand, and how to keep your history going in your next tool.

6 min
Read guide
ImplementationIntermediate

Replacing Stackify Retrace Agents: Java, PHP, Node.js, Python and Ruby

Swap the Stackify Retrace agent for the Atatus agent in Java, PHP, Node.js, Python and Ruby apps, one app at a time, so nothing goes unmonitored before Retrace shuts down on March 31, 2027.

9 min
Read guide
ImplementationIntermediate

Moving from Stackify Retrace to OpenTelemetry

BMC recommends OpenTelemetry as the standard to replace the Stackify Retrace agents. What OpenTelemetry captures compared with the Retrace profilers, how to send the same data to two tools while you choose, and how to send it to Atatus.

9 min
Read guide