Table of Contents
PU, memory use, latency, network bandwidth. These are just some of the monitoring metrics businesses analyze for security and performance. But successful data-driven organizations delve deeper than this. These companies probe millions of real-time metrics for unexpected insights and predict outcomes weeks, months, and years into the future. ELK helps them do this. It's a data analytics platform from open-source developer Elastic.
ELK touts itself as an essential piece of tech for log analytics that lets users "securely and reliably search, analyze, and visualize" data. But does it work? And is there a better alternative?
In this ELK review, we've listed the pros, cons, and features of this open-source framework for monitoring. Learn what customers think and discover alternative solutions from MetricFire.
MetricFire is a monitoring platform built on open source monitoring technology such as Graphite, Grafana, and Prometheus. You can build Grafana dashboards based on your hosted Prometheus and Graphite data stores directly in the app. Check out our free trial, or sign up for a demo and talk to the team directly.
What is ELK?
ELK — an acronym of Elasticsearch, Logstash, and Kibana — is an open-source project for log analytics developed by Amsterdam, Netherlands-based company Elastic N.V. Previously known as the "ELK Stack," and now referred to as the "Elastic Stack," ELK consists of the three products in its acronym and a fourth product called Beats.
- Elasticsearch: Java-based search engine built on the Lucene library. It's used for logging, monitoring, searching, and analysis.
- Logstash: Log management tool that collects, processes, and forwards data (like events and log messages). It's the pipeline that sends data from multiple sources to Elasticsearch in the right format.
- Kibana: Open-source visualization dashboard that lets users create maps and charts on top of volumes of data. It's the interface that lets users see their "Elastic Stack."
- Beats: Data shippers that send operational data to Logstash.
If you like Elasticsearch products, but prefer Grafana, you can always import your Elasticsearch data into your Grafana dashboard. Check out the tutorial here.
- Prometheus is a powerful open-source monitoring system built on a time-series database that lets users receive and store data, create graphs, set up alerts, and execute queries. Users can upgrade to Hosted Prometheus and store data for up to 2 years. (Open-source Prometheus has a retention period of 15 days.)
- Graphite is a push-based open-source software that facilitates monitoring tasks in various contexts. It collects data and creates visualizations (presented on beautiful Grafana dashboards) that help businesses monitor networks, servers, applications, and other tasks. Users can set up alerts for deeper monitoring insights. Hosted Graphite packages come with Hosted Grafana.
ELK Features: How Do They Compare to MetricFire?
ELK presents data visualizations on powerful dashboards via Kibana. (The data itself comes from Elasticsearch.) Kibana executes various data aggregation types, although there are field type/configuration limitations. For example, aggregations for histograms (displays of data that use bars of different heights) only work on integers, while "min and max metric" aggregations only work on date and number fields. Only advanced users will notice these limitations, however, as there are so many different aggregation types.
MetricFire's visualizations are just as powerful as those on ELK, with feature-rich Grafana dashboards that utilize an interactive, intuitive UI. (Grafana dashboards come with all MetricFire packages). There are few limitations with Grafana; users can add rows and panels, customize graphs, and zoom into visualizations with ease. This lets users better understand all kinds of data sets, including server delays, sensor alert times, rebuild times, and device delays.
ELK's "Graph" feature, part of Elasticsearch, finds meaningful relationships between various data sets — real-time data availability, indexing, distribution query execution, etc. — for deeper, clearer monitoring insights. Users can visualize these relationships through Kibana.
MetricFire's Prometheus, however, provides users with a multi-dimensional data model that connects to target endpoints via HTTP. Users can define tags/names with a unique "time series" of data. Graphite, the other product from MetricFire, works with system statistics collection daemons StatsD and collectd to pull metrics from the user's infrastructure and stores these metrics in a database called Whisper. Users can then make data connections that power business-making decisions.
ELK works with various external tools to optimize Elasticsearch's logging, monitoring, and searching capabilities. These integrations include WordPress, Drupal, and Tiki.
MetricFire also has numerous integrations, providing support for AWS, Heroku, and logging tools. Users can execute and modify integrations with various APIs and create notifications for actionable events via email, Slack, and other services.
ELK User Support
The level of support users receive depends on their subscription level (more on that in the next section.) All users can submit a support ticket (ELK calls these "cases"), but only users with more expensive subscription levels can access telephone support. However, users still need to submit a case before they can phone ELK. (ELK doesn't provide its phone number otherwise.)
MetricFire, on the other hand, provides the same level of support for all users, with both email, phone, and video call (via a service called appear.in) options. You can contact MetricFire 24 hours a day, seven days a week.
Is MetricFire Cheaper Than ELK?
ELK bases its pricing on resources used, but it's extremely complicated, with far too many subscription levels to list in this review. There are basic and enterprise-level tiers for self-managed cloud services, 4-5 additional tiers for managed cloud services, and a handful of other tiers for the Elastic Stack itself. To complicate things further, ELK doesn't publish prices on its website — Elastic asks you to contact them — which makes it difficult for comparison purposes.
MetricFire, without doubt, has a more simplified pricing structure, with the most expensive package (Premium Annual for large teams who need expansive metrics and scaling) costing $3,849 per month. Newbies can try free monitoring for 14 days.
ELK Customer Reviews
Elasticsearch, ELK's flagship product, currently holds a 4.6/5 rating on the software review website G2 (based on 51 reviews).
"Elasticsearch is NO-SQL search engine full of features and support document format. It supports REST API and makes it easy for users to send a search request using any language (back-end)," says one computer science student who works for a large enterprise.
Other users point out Elasticsearch's flaws. "When doing many search requests, the server shuts down sometimes," says a graduate student researcher working for a mid-market company. "Also, the speed of accessing the data, even when using the bulk API, is low."
MetricFire has fewer customer reviews than ELK on G2. However, customers enjoy MetricFire's products. "Hosted Graphite makes it easy for me to keep track of my remote devices used by customers," says a clinical account manager in the healthcare sector. "I can't always be on-site, so this allows me easy tracking of device details."
ELK's open-source products provide users with streamlined visualizations, data relationships, and integrations for more effective monitoring. However, it has a complicated pricing structure and user support limitations. MetricFire, on the other hand, provides flexible customer support and simple pricing, as well as deep insights into all monitoring environments.