Hue Flies High at Goibibo
Our thanks to Amar Parkash, a Software Developer at Goibibo, a leading travel portal in India, for the enthusiastic support of Hue you’ll read below.
At Goibibo, we use Hue in our production environment. I came across Hue while looking for a near real-time log search tool and got to know about Cloudera Search and the interface provided by Hue. I tried it on my machine and was really impressed by the UI it provides for Apache Hive, Apache Pig, HDFS, job browser, and basically everything in the Big Data domain. We immediately deployed Hue in production, and that has been one of the best decisions we have ever made for our data platform at Goibibo.
Following are just the some of the problems Hue has solved for us:
- Cloudera Search: As I mentioned above, search was the first use case we started using Hue for. We have integrated Cloudera Search with Flume NG. At present, developers are actively using Hue (for search) and it has helped them save a lot of time. Previously people used grep or other shell commands which was painful and time consuming.
- Hive: This has been the use case where Hue has been a boon for us in data platform. Previously, businesspeople would bring us their queries, and then we would give them the output by running the Hive queries in terminal. Thanks to Hue, we have created saved queries that businesspeople can directly run by specifying the parameters. Also, the option to download the output as csv and xls is really amazing. Previously, it was a big hassle to transfer the files.
- Cloudera Impala: We were not using Impala much until December 2013. Since Impala is integrated with Hue, I thought of trying it. I quickly set up some impala daemons using Cloudera Manager and I was amazed to see the performance improvement in the execution of queries as compared to Hive. Since then, we have been actively using Impala, and as mentioned previously, it has been a big timesaver for our businesspeople as well. Now we are running some queries that we never thought of running in Hive due to poor performance.
- File Browser: I can’t remember when the last time I used a
hadoop fs -lscommand. The file browser is just too good.
- Job Browser: The job browser comes in very handy in cases where Hive queries take too much time or get stuck.
- User Admin: The user admin panel is also very helpful. There are some cases where we don’t want to give some particular access to every user and the user admin panel does that precisely.
I would give Hue a 5-out-of-5 rating and recommend it to anyone working on a Hadoop cluster!