Mon­i­tor­ing Docker In­fra­struc­ture

OpenSource For You - - CONTENTS -

Docker brings in porta­bil­ity to tech­nol­ogy-in­ten­sive soft­ware devel­op­ment en­vi­ron­ments. As a re­sult, it has be­come quite popular among de­vel­op­ers to demon­strate a proof-of-con­cept, tech­nol­ogy pre­view re­lease through the use of Docker con­tain­ers.

At the same time, Docker runs UNIX pro­cesses on Linux with strong guar­an­tees of iso­la­tion, thanks to Linux ker­nel tech­nolo­gies like names­paces and con­trol groups (cgroups). This has dis­rupted the en­tire de­ploy­ment and op­er­a­tions space and proven Docker to be a great build­ing block for au­tomat­ing soft­ware in­fra­struc­ture, namely, large-scale Web de­ploy­ments, data­bases, pri­vate Plat­forms as a Ser­vice (PaaS), con­tin­u­ous de­ploy­ment sys­tems and ser­vice-ori­ented in­fra­struc­tures (SOA), among oth­ers.

In sim­ple terms, the Docker run­time it­self is a con­tainer based PaaS plat­form. ef­fi­ciency. How­ever, con­tainer­i­sa­tion dis­rupted the en­tire game with fru­gal sys­tem re­source utilisation, pow­ered by op­er­at­ing sys­tem soft­ware ca­pa­bil­i­ties, thanks to names­paces and cgroups.

As a re­sult, one can run thou­sands of ‘dock­erised’ por­ta­ble apps on com­mod­ity hard­ware, driv­ing for­ward sig­nif­i­cant ef­fi­ciency and scale on the same plat­form that ear­lier strug­gled un­der the vir­tu­al­i­sa­tion work­load of tens of op­er­at­ing sys­tems.

Given this flex­i­bil­ity, it is en­tirely pos­si­ble that a bunch of re­source-hun­gry ap­pli­ca­tions can sig­nif­i­cantly re­duce the per­for­mance of the other ‘dock­erised’ apps. The ob­jec­tive of mon­i­tor­ing is to pre­cisely ad­dress this im­por­tant case by col­lect­ing op­er­at­ing sys­tem level met­rics and hav­ing a mech­a­nism to man­age the apps to en­sure scal­a­bil­ity. This

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.