text
stringlengths
0
1.8k
[2980.30 --> 2983.70] of the CD processes, including the logs,
[2984.34 --> 2985.72] as something critical.
[2985.96 --> 2988.38] And we think that using open telemetry,
[2988.50 --> 2991.44] it will be easier than ever to route
[2991.44 --> 2993.96] all your audit trail of your release
[2993.96 --> 2996.42] processes, build process of what you
[2996.42 --> 2998.38] ship in production, to route them
[2998.38 --> 3001.38] directly in this very secure, long-term,
[3001.54 --> 3004.02] cost-effective storage, being your logs
[3004.02 --> 3004.80] management system.
[3004.80 --> 3007.36] It could be maybe an S3 bucket or maybe,
[3007.58 --> 3009.66] let's say, your Splunk, Elastic, or you
[3009.66 --> 3011.26] name it, a long-term storage.
[3012.04 --> 3013.88] So this is what comes to my mind.
[3013.98 --> 3015.50] And then there are some other
[3015.50 --> 3018.20] requirements for the CICD companies,
[3018.38 --> 3020.46] but I am less involved in this at the
[3020.46 --> 3020.72] moment.
[3020.72 --> 3022.24] How do you think about supply and chain
[3022.24 --> 3024.76] security within the CICD space, Oleg?
[3024.76 --> 3026.42] I definitely support this topic.
[3026.58 --> 3027.92] It's very important.
[3028.38 --> 3031.84] And when SolarWinds was announced one year
[3031.84 --> 3034.02] ago, we actually had a Jenkins
[3034.02 --> 3035.66] governance board meeting and then
[3035.66 --> 3037.18] discussion at the contributor summit.
[3037.18 --> 3039.82] And we decided to prioritize supply and
[3039.82 --> 3041.70] chain security as one of the major
[3041.70 --> 3043.38] topics for this year for Jenkins
[3043.38 --> 3043.94] community.
[3044.64 --> 3046.36] And if you have seen, there is a lot of
[3046.36 --> 3047.66] activities on this front.
[3047.78 --> 3049.30] For example, dependency updates.
[3049.54 --> 3051.62] We invest quite a lot in tooling,
[3052.10 --> 3054.24] in dependency scanning, in bills of
[3054.24 --> 3054.94] materials.
[3055.28 --> 3057.26] So currently we can produce S-bombs for
[3057.26 --> 3058.48] companies if needed.
[3059.54 --> 3061.44] And indeed, this is important.
[3061.60 --> 3064.10] And it's important for us because we are
[3064.10 --> 3065.50] a second-level supplier.
[3065.50 --> 3067.24] We depend on so many libraries.
[3067.64 --> 3069.22] We need to verify them, but we also
[3069.22 --> 3071.64] need to provide a good level of trust
[3071.64 --> 3074.20] so that users of Jenkins and of our
[3074.20 --> 3076.02] systems can safely deliver their
[3076.02 --> 3076.58] software.
[3077.12 --> 3078.52] Something that comes to my mind here
[3078.52 --> 3080.76] that I touched when I was working on
[3080.76 --> 3082.98] CI and that I see also now that I work
[3082.98 --> 3086.16] on observability is the importance of
[3086.16 --> 3088.62] capturing the right information in the
[3088.62 --> 3089.36] bill of material.
[3090.12 --> 3091.86] And I think it's also an incremental
[3091.86 --> 3092.38] journey.
[3092.82 --> 3094.84] First, you build on your Docker
[3094.84 --> 3096.30] environment, but if you don't capture
[3096.30 --> 3100.12] exactly the hash of the Docker image that
[3100.12 --> 3103.18] was used to run your build, it's too late.
[3103.26 --> 3105.26] You will not be able to re-understand it
[3105.26 --> 3106.16] 12 months later.
[3106.74 --> 3108.48] And so I think there is an incremental
[3108.48 --> 3109.04] journey.
[3109.38 --> 3111.62] It's a continuous exercise to verify that
[3111.62 --> 3114.82] the data you capture in your build are good
[3114.82 --> 3117.12] enough to understand what actually happened.
[3117.12 --> 3119.28] You mentioned the problem is the usage of
[3119.28 --> 3119.98] cache system.
[3119.98 --> 3123.48] Do I capture all the details to understand
[3123.48 --> 3125.26] what artifact was retrieved from my
[3125.26 --> 3126.10] caching system?
[3126.26 --> 3127.08] Have I been poisoned?
[3127.94 --> 3130.40] And this is a never-ending exercise in
[3130.40 --> 3132.64] some ways to always capture the right
[3132.64 --> 3133.94] metadata on your build.
[3134.36 --> 3136.54] Is Captain Obvious involved in any of
[3136.54 --> 3136.90] this, Oleg?
[3137.04 --> 3139.50] Yes and no, because I'm currently building
[3139.50 --> 3142.46] a prototype which integrates Jenkins,
[3142.84 --> 3144.72] OpenTelemetry, and Captain.
[3144.72 --> 3147.40] But for me, the main objective is to actually
[3147.40 --> 3149.48] expose more information about quality
[3149.48 --> 3149.92] gates.
[3150.20 --> 3152.66] So when we deliver software, we can verify
[3152.66 --> 3155.42] that all items are basically delivered
[3155.42 --> 3157.58] with all matching criteria.
[3157.86 --> 3160.10] So currently, Captain is mostly built around
[3160.10 --> 3161.96] cloud events, which is probably a topic for
[3161.96 --> 3162.72] separate discussion.
[3163.36 --> 3165.94] Captain exposes OpenTelemetry metrics on
[3165.94 --> 3168.48] its own, so you can understand what happens
[3168.48 --> 3170.34] inside Captain when you analyze, for example,
[3170.50 --> 3171.66] quality gates, etc.
[3171.66 --> 3174.42] But it would also be great to have integration
[3174.42 --> 3175.40] in other directions.
[3175.72 --> 3179.18] So when CI-CD systems supply information about
[3179.18 --> 3182.38] the status, metrics, and especially all
[3182.38 --> 3185.22] deployment parameters, tools like Captain,
[3185.40 --> 3187.66] so that they can make decisions whether the
[3187.66 --> 3190.40] system is compliant with the expectations