September 2015 Summaries

Graphite is a popular real-time metrics storage and graphing tool that allows storing numeric time-series data and graphing it on demand, but its architecture has limitations when scaling to thousands of metrics per second. To address these scalability issues, the graphite-ng project was created in Go, which outperforms vanilla graphite installations and can handle millions of metrics per second with features like data validation, metric series combination, and faster aggregation. InfluxDB is another scalable solution that aims to solve problems with Whisper, its default storage engine, and offers a more scalable backend database for storing time-series data efficiently, with high availability, clustering, and horizontal scalability in future releases.

Sending sensor data to InfluxDB from an Arduino Uno can be achieved using the UDP protocol, which requires minimal processing and networking overhead. The InfluxDB line protocol provides a compact way to write data over HTTP or UDP, with points specified as key-value pairs separated by spaces. To enable UDP on InfluxDB, modify the configuration file to specify the database and bind address. A sample code example uses the TMP36 temperature sensor to poll once per second and send the value to the specified InfluxDB host via UDP. Once data is written to InfluxDB, queries can be issued using `curl` or other tools to retrieve it, and troubleshooting tips are provided in case of issues with data collection.

InfluxDB Week in Review - September 28, 2015` InfluxDB was used by Didi Kuaidi, a Chinese ride-hailing company with over 100 million users, to manage its thousands of servers and maintain service availability. InfluxDB also powered KNXMonitor, a cloud-based remote debugging and monitoring system for KNX installations. The influxdb-java client library now supports InfluxDB 0.9.x, and users can calculate the derivative of a sum using continuous queries. To celebrate its community support, InfluxDB offered free hoodies to companies with production-ready apps or products built on top of InfluxDB.

Telegraf 0.1.9 has been released with several new features, including clustering support for InfluxDB clusters, filtering of telegraf outputs on the CLI and in config files, Makefile GOBIN support, AMQP output plugin, OpenTSDB output plugin, retry failed server connections, and a port tag to the Apache plugin. The release also includes various breaking changes, such as refactored CPU collection and memory plugins, with some measurements renamed for consistency and others removed from output in default config. Additionally, 11 bug fixes have been included in the release.

InfluxDB's SHOW STATS command provides runtime statistics and diagnostics meta queries to monitor the system, including memory usage, HTTP requests, and input performance metrics such as Graphite and openTSDB. The internal database stores statistical data that can be analyzed using InfluxQL, allowing for long-term performance analysis and visualization through tools like Chronograf. Additionally, cluster-level statistics are available for the whole cluster, while diagnostic information is stored separately from the internal database and includes system details such as build version and uptime.

InfluxDB has recently released version 0.9.4 with new features such as the top function and order by time desc, allowing users to better analyze their data. The company is also presenting at upcoming meetups in San Francisco, Denver, and Chicago, where they will share knowledge on using InfluxDB for SQL DBAs and other topics. Users can get answers to common questions about aggregating data over specific time periods and shutting down the database gracefully through official tutorials and documentation. Additionally, InfluxDB is seeking submissions from companies using their product in production, offering a "thank you" gift of an InfluxDB hoodie and stickers for featured entries.

Telegraf is a daemon that collects metrics from systems, services, and third-party APIs, and outputs them to InfluxDB and Kafka. It is plugin-driven, allowing for easy extension and modification of data collection and output. Telegraf was originally built for InfluxDB but has evolved to support other data sinks like Kafka, Datadog, and OpenTSDB. The guide covers installing and configuring Telegraf, running it on a system, viewing its data in Kafka and InfluxDB, and exploring additional features and plugins.

InfluxDB 0.9.4 has been released with several notable features including the top aggregate function, order by time descending, and backreferencing measurement names in continuous queries. The new release also includes instrumentation information about InfluxDB through `show stats` and `show diagnostics` queries. Additionally, this version is built with Go 1.5, which may result in slightly different performance characteristics. The next priority for the project is improving write throughput and clustering capabilities with a new storage engine and bug fixes. The release also includes support for a free trial of managed InfluxDB servers in the cloud and expert professional services for help with InfluxDB projects.