Telemetry¶

Starting with version v1.1.1 of StripePy, we introduced support for telemetry collection.

This page outlines what information we are collecting and why. Furthermore, we provide instructions on how telemetry collection can be disabled.

What information is being collected¶

stripepy is instrumented to collect general information about stripepy itself and the system where it is being run.

We do not collect any sensitive information that could be used to identify our users, the machine where stripepy is being run, or the datasets processed by stripepy.

This is the data we are collecting:

Information on how stripepy was installed (i.e., the package version and third-party dependency versions).
Information on the system where stripepy is being run (i.e., operating system, processor architecture, and Python version).
How stripepy is being invoked (i.e., the subcommand, input file format and parameters).
Information about stripepy execution (i.e., when it was launched, how long the command took to finish, and whether the command terminated with an error).

The following table shows an example of the telemetry collected when running stripepy call ENCFF993FGR.hic 10000 -p 8:

Telemetry information collected when running `stripepy call`¶
Field	Value
dependency.h5py.version	3.14.0
dependency.hictkpy.version	1.3.0
dependency.numpy.version	2.3.1
dependency.packaging.version	25.0
dependency.pandas.version	2.3.1
dependency.scipy.version	1.16.0
dependency.structlog.version	25.4.0
duration_ms	103074.812875
host.arch	x86_64
library.name	stripepy
meta.signal_type	trace
name	call
os.type	linux
os.version	6.11.0-1015-azure
params.constrain_heights	false
params.contact_map_format	mcool
params.contact_map_raw_interactions	true
params.contact_map_resolution	20000
params.genomic_belt	5000000
params.glob_pers_min	0.04
params.k	3
params.loc_pers_min	0.33
params.loc_trend_min	0.25
params.max_width	100000
params.nproc	4
process.runtime.description	GCC 12.2.0
process.runtime.name	CPython
process.runtime.version	3.13.5
Sample Rate	1
service.name	stripepy
service.version	1.1.1.dev73+g2e13fec
span.kind	internal
span.num_events	0
span.num_links	0
status_code	1
telemetry.sdk.language	python
telemetry.sdk.name	opentelemetry
telemetry.sdk.version	1.34.1
trace.span_id	4b6f5534c8aec420
trace.trace_id	30b4804d781557f552d15dec8270fca8
type	internal

Why are we collecting this information?¶

There are two main motivations behind our decision to start collecting telemetry data:

To get an idea of how big our user base is: this will help us, among other things, to secure funding to maintain stripepy in the future.
To better understand which of the functionalities offered by stripepy are most used by our users: we intend to use this information to help us decide which features we should focus our development efforts on.

How is telemetry information processed and stored¶

Telemetry is sent to an OpenTelemetry collector running on a virtual server hosted on the Norwegian Research and Education Cloud (NREC).

The virtual server and collector are managed by us, and traffic between stripepy and the collector is encrypted.

The collector processes incoming data continuously and forwards it to a dashboard for data analytics and a backup solution (both services are hosted in Europe). Communication between the collector, dashboard, and backup site is also encrypted. Data stored by the dashboard and backup site is encrypted at rest.

The analytics dashboard keeps telemetry data for up to 60 days, while the backup site is currently set up to store telemetry data indefinitely (although this may change in the future).

How to disable telemetry collection¶

To disable telemetry collection, simply define the STRIPEPY_NO_TELEMETRY environment variable before launching stripepy (e.g., STRIPEPY_NO_TELEMETRY=1 stripepy download)

Where can I find the code used for telemetry collection?¶

All code concerning telemetry collection is defined in file src/stripepy/cli/telemetry.py.