I have a simple setup to get telemetry off an energy meter and graph it.
Meter > MQTT > RabbitMQ > Python listener > QuestDB > Grafana.
I got it all working on a single EC2 instance running as services. I then rebuilt in a new instance using Docker - separate container for RabbitMQ, QuestDB, python script and Grafana.
I can get the script to run from inside its own container, consuming from rabbitmq in another.
But it wont write to the quest db database in another container.
I can ping from my "listener" container to "rabbitmq" and "questdb" containers (from inside "listener"), the python library for RabbitMQ works fine (pika) but now the questdb "Sender" call I make that used to work just fine when the python was running from the host to the containers now won't work - gives
AttributeError: type object 'questdb.ingress.Sender' has no attribute 'from_conf'
And it’s not giving a connection failure error, it’s just puking instantly.
I can successfully write to the questdb db using that same python script from a different EC2 instance.
It seems like the Sender func does not like talking from container to container inside the same host.
Code:
with Sender.from_conf(conf) as sender:
sender.row(
'Sensu',
columns={'DeviceID': DeviceID, 'Voltage_rms': payload[0], 'Current_rms': payload[1], 'Power_real':payload[2], 'Import_energy': payload[3], 'Export_energy': payload[4], 'Frequency': payload[5], 'PowerFactor': payload[6], 'ListenerTimestamp': ListenerTimestamp, "MessageCount": MessageCount},
at=TimestampNanos.now())
Using
conf = f'http::addr=localhost:9000;' << every permutation of this, localhost, container network name, internal IP address, with and without auth.
Docker-compose.yaml:
name: sensu
services:
rabbitmq:
image: rabbitmq:3-management
container_name: rabbitmq
hostname: rabbitmq-server
restart: unless-stopped
ports:
- 5672:5672
- 15672:15672
- 1883:1883
environment:
- COMPOSE_PROJECT_NAME=sensu
- RABBITMQ_DEFAULT_PASS=sensupass
- RABBITMQ_DEFAULT_USER=sensu-mqtt
command: "/bin/bash -c \"rabbitmq-plugins enable --offline rabbitmq_mqtt; rabbitmq-server\""
volumes:
- ~/.docker-conf/rabbitmq/data/:/var/lib/rabbitmq/
- ~/.docker-conf/rabbitmq/log/:/var/log/rabbitmq
networks:
- sensu_net
questdb:
image: questdb/questdb
container_name: questdb
hostname: questdb-server
restart: unless-stopped
ports:
- 9000:9000
environment:
- COMPOSE_PROJECT_NAME=sensu
volumes:
- ~/questdb_data/questdb:/root/.questdb/db
networks:
- sensu_net
networks:
sensu_net:
name: sensu_net
driver: bridge
I have a simple setup to get telemetry off an energy meter and graph it.
Meter > MQTT > RabbitMQ > Python listener > QuestDB > Grafana.
I got it all working on a single EC2 instance running as services. I then rebuilt in a new instance using Docker - separate container for RabbitMQ, QuestDB, python script and Grafana.
I can get the script to run from inside its own container, consuming from rabbitmq in another.
But it wont write to the quest db database in another container.
I can ping from my "listener" container to "rabbitmq" and "questdb" containers (from inside "listener"), the python library for RabbitMQ works fine (pika) but now the questdb "Sender" call I make that used to work just fine when the python was running from the host to the containers now won't work - gives
AttributeError: type object 'questdb.ingress.Sender' has no attribute 'from_conf'
And it’s not giving a connection failure error, it’s just puking instantly.
I can successfully write to the questdb db using that same python script from a different EC2 instance.
It seems like the Sender func does not like talking from container to container inside the same host.
Code:
with Sender.from_conf(conf) as sender:
sender.row(
'Sensu',
columns={'DeviceID': DeviceID, 'Voltage_rms': payload[0], 'Current_rms': payload[1], 'Power_real':payload[2], 'Import_energy': payload[3], 'Export_energy': payload[4], 'Frequency': payload[5], 'PowerFactor': payload[6], 'ListenerTimestamp': ListenerTimestamp, "MessageCount": MessageCount},
at=TimestampNanos.now())
Using
conf = f'http::addr=localhost:9000;' << every permutation of this, localhost, container network name, internal IP address, with and without auth.
Docker-compose.yaml:
name: sensu
services:
rabbitmq:
image: rabbitmq:3-management
container_name: rabbitmq
hostname: rabbitmq-server
restart: unless-stopped
ports:
- 5672:5672
- 15672:15672
- 1883:1883
environment:
- COMPOSE_PROJECT_NAME=sensu
- RABBITMQ_DEFAULT_PASS=sensupass
- RABBITMQ_DEFAULT_USER=sensu-mqtt
command: "/bin/bash -c \"rabbitmq-plugins enable --offline rabbitmq_mqtt; rabbitmq-server\""
volumes:
- ~/.docker-conf/rabbitmq/data/:/var/lib/rabbitmq/
- ~/.docker-conf/rabbitmq/log/:/var/log/rabbitmq
networks:
- sensu_net
questdb:
image: questdb/questdb
container_name: questdb
hostname: questdb-server
restart: unless-stopped
ports:
- 9000:9000
environment:
- COMPOSE_PROJECT_NAME=sensu
volumes:
- ~/questdb_data/questdb:/root/.questdb/db
networks:
- sensu_net
networks:
sensu_net:
name: sensu_net
driver: bridge
- 1 Which version of the python client are you using? from_conf was added about a year ago in version 2.0.0 py-questdb-client.readthedocs.io/en/latest/changelog.html#id7 – Javier Ramirez Commented Feb 17 at 14:54
- Should be latest... installing using "pip install -U questdb". – d0dja Commented Feb 17 at 14:58
- 1 Could you please verify with pip list inside the container running the client? – Javier Ramirez Commented Feb 17 at 15:39
- WIll do. FWIW this is the Dockerfile (using compose for questdb and rabbit as they're done; still working on the listener. – d0dja Commented Feb 17 at 15:50
- Dockerfile: FROM python:3.7-alpine COPY . / RUN pip3 install pika questdb WORKDIR /listener pip install --upgrade pip COPY telemetry_insert.py /listener CMD ["python3", "telemetry_insert.py"] – d0dja Commented Feb 17 at 15:51
1 Answer
Reset to default 0From your comment, this is a minimal Dockerfile you are using:
FROM python:3.7-alpine
COPY . /
RUN pip3 install -U pika
RUN pip3 install -U questdb
Building this, creating a test container, and checking the package version for questdb:
❯ docker run --rm -it --entrypoint /bin/sh test
/ # pip list
Package Version
---------- -------
pika 1.3.2
pip 23.0.1
questdb 1.1.0
setuptools 57.5.0
wheel 0.41.2
Since from_conf was introduced in questdb==2.0.0, this is the result of your error. To fix this, you will need to install a newer version of questdb. To do this, change your Dockerfile to a more recent Python version. For example:
FROM python:3.12-alpine
This gives you a more recent questdb version:
❯ docker run --rm -it --entrypoint /bin/sh test
/ # pip list
Package Version
------- -------
pika 1.3.2
pip 24.3.1
questdb 2.0.3
For future reference, navigate to the PyPi package link and scroll down to find the required Python version (Requires: Python >=Y.XZ) for your package: https://pypi./project/questdb/
- Credit to @Javier Ramirez for finding the package release notes for v2.0.0